Housing Economics and Public Policy: Essays in Honour of Duncan Maclennan

Housing Economics and Public Policy Edited by Tony O’Sullivan Head of Planning Communities Scotland and Kenneth Gibb R...

Author: Tony O’Sullivan | Kenneth Gibb

432 downloads 799 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Housing Economics and Public Policy Edited by

Tony O’Sullivan Head of Planning Communities Scotland and

Kenneth Gibb Reader Department of Urban Studies University of Glasgow

Essays in honour of Duncan Maclennan

Blackwell Science

Housing Economics and Public Policy

Housing Economics and Public Policy Edited by

Tony O’Sullivan Head of Planning Communities Scotland and

Kenneth Gibb Reader Department of Urban Studies University of Glasgow

Essays in honour of Duncan Maclennan

Blackwell Science

© 2003 by Blackwell Science Ltd, a Blackwell Publishing company Editorial Ofﬁces: Blackwell Science Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK Tel: +44 (0) 1865 776868 Blackwell Publishing Inc., 350 Main Street, Malden, MA 02148-5020, USA Tel: +1 781 388 8250 Blackwell Science Asia Pty, 550 Swanston Street, Carlton, Victoria 3053, Australia Tel: +61 (0)3 8359 1011 The right of the Author to be identiﬁed as the Author of this Work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. First published 2003 by Blackwell Science Ltd Reprinted 2006 ISBN-10: 0-632-06461-7 ISBN-13: 978-0-632-06461-8 Library of Congress Cataloging-in-Publication Data is available A catalogue record for this title is available from the British Library Set in 10/13pt Trump Mediaeval by Sparks, Oxford – www.sparks.co.uk Printed and bound in Great Britain by Marston Book Services, Oxford The publisher’s policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp processed using acid-free and elementary chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. For further information on Blackwell Publishing, visit our website: www.blackwellpublishing.com

Real Estate Issues Series Managing Editors Stephen Brown RICS Foundation John Henneberry Department of Town & Regional Planning, University of Shefﬁeld James Shilling Department of Real Estate and Urban Land Economics, University of Wisconsin – Madison

Real Estate Issues is a book series presenting the latest thinking into how real estate markets operate. It is inclusive in nature, drawing both upon established techniques for real estate market analysis and on those from other academic disciplines. It embraces a comparative approach, allowing best practice to be put forward and tested for its applicability and relevance to the understanding of new situations. It does not impose solutions, but provides a means by which solutions can be found. Real Estate Issues does not make any presumptions as to the signiﬁcance of real estate markets, but presents the real signiﬁcance of the operation of these markets.

Books in this series Guy & Henneberry Development and Developers Adams & Watkins Greenﬁelds, Brownﬁelds and Housing Development O’Sullivan & Gibb Housing Economics and Public Policy Couch, Fraser & Percy Urban Regeneration in Europe Stephens Housing Finance and Owner-occupation Brown & Jaffe Real Estate Investment Seabrooke & How International Real Estate Allen & Barlow Housing in Southern Europe Ball Markets and Institutions in Real Estate and Construction

For Amanda, Amy, Gemma, and Angela

Contents Contributors Preface Acknowledgements 1 Introduction Tony O’Sullivan & Kenneth Gibb Perspectives on theory Policy and practice Contributions to this book Recurrent themes and key issues for the future 2 Urban Housing Models Kenneth Gibb Introduction Urban residential trade-off models Filtering in urban housing markets Simulation models of urban housing Conclusions

xiii xix xxi 1 1 4 9 18 22 22 23 27 30 35

3 Segmentation, Adjustment and Disequilibrium Yong Tu Introduction The causes and consequences of housing market segmentation A model of housing market disequilibrium Urban housing sub-market identiﬁcation Conclusions

38

4 Transactions Costs and Housing Markets John M. Quigley Introduction A taxonomy A simple model of transactions costs Some policy implications

56

5 Hedonic Pricing Models: a Selective and Applied Review Stephen Malpezzi Introduction What is a hedonic price index? Repeat sales models The roots of hedonic price models

38 38 42 51 54

56 57 61 63 67 67 68 71 73

x

Contents

Conceptual issues in hedonic modelling Speciﬁcation issues Hedonic modelling: the current position Examples of applications Concluding thoughts 6 Housing, Random Walks, Complexity and the Macroeconomy Geoffrey Meen Introduction The relationship between housing and the macroeconomy Are house prices predictable? Conclusions

75 76 82 83 85 90 90 92 98 107

7 Taxation, Subsidies and Housing Markets Gavin A. Wood Introduction Distributional issues Tax arbitrage Taxation, inefﬁciency and market processes Concluding comments

110

8 The Economics of Social Housing Christine M.E. Whitehead Introduction: the role of economic analysis in social housing Efﬁciency reasons for social housing provision Housing as a means of redistribution The governance of housing Conclusions

135

9 Neighbourhood Dynamics and Housing Markets George C. Galster Introduction Deﬁning neighbourhood Idiosyncrasies of neighbourhood How neighbourhoods come to be Changes in neighbourhood Implications for neighbourhood regeneration policy Summary and conclusion

153

10 Access to Home Ownership in the United States: the Impact of Changing Perspectives on Constraints to Tenure Choice Joseph Gyourko Introduction

110 112 118 124 129

135 138 143 148 151

153 153 156 158 160 167 167

172 172

Contents

Constraints to tenure choice – theory and empirical evidence Wealth constraints and racial differences in ownership in the United States Implications for the affordability debate Summary and conclusion 11 Planning Regulation and Housing Supply in a Market System Glen Bramley Introduction Researching housing supply Planning and regulation of land use Setting up an inter-urban panel model Model results Conclusions 12 Economics and Housing Planning Tony O’Sullivan Introduction The policy and institutional framework Housing plans in practice The land use planning system and housing planning The roots of the problem Conclusions 13 The Right to Buy in Britain Nick Williams Introduction Policy objectives The evidence base for the 1980 legislation Subsequent research on the statutory Right to Buy The relationship between policy and research The modernised or strategic Right to Buy in Scotland Conclusions 14 The Political Economy of Housing Research David Donnison & Mark Stephens Introduction Housing research resumes after World War II Recent research on housing A changing environment for research The housing research industry In conclusion

xi

174 178 182 188 193 193 195 196 201 206 216 218 218 219 222 225 226 232 235 235 236 237 238 241 242 245 248 248 249 251 255 262 266

xii

Contents

15 Policy and Academia: an Assessment Richard Best Introduction Separate worlds? Crossing the divide Making the connections In conclusion

268 268 268 271 275 281

References Index

283 319

Contributors Richard Best ([email protected]) Director of the Joseph Rowntree Foundation in York and its housing association, the Joseph Rowntree Housing Trust. He was previously Director of the National Federation of Housing Associations and before that Director of the British Churches Housing Trust. He has been involved in housing policy issues for some thirty years, including serving on many government bodies, task forces, Ministerial Sounding Boards, etc. He became an independent life peer in 2001.

Glen Bramley ([email protected]) Professor of Planning & Housing at Heriot-Watt University in Edinburgh, where he leads a substantial research programme in housing and urban studies. Recent work is focused particularly on housing need and areas of low demand for housing, ﬂows of funds to local areas, planning for new housing and the impact of planning and infrastructure on city competitiveness. He directs a Centre for Research into Socially Inclusive Services (CRSIS). His publications include Planning, the Market and Private Housebuilding (UCL Press 1995), Equalization Grants and Local Expenditure Needs (Avebury 1990) and Analysing Social Policy (Blackwell 1986), as well as numerous articles in journals such as Urban Studies, Housing Studies, Environment and Planning, Regional Studies, Policy & Politics.

David Donnison ([email protected]) Honorary research fellow and professor emeritus in the Department of Urban Studies, University of Glasgow. He was previously Glasgow’s professor of Town and Regional Planning and, before that, chair of the Supplementary Beneﬁts Commission and Director of the Centre for Environmental Studies. His latest book is Policies for a Just Society (Macmillan 1998), and his more recent writing has dealt mainly with poverty, and the evolution of social policies since World War II.

George Galster ([email protected]) Clarence Hilberry Professor of Urban Affairs at the College of Urban, Labor, and Metropolitan Affairs, Wayne State University, Detroit, USA. He earned his Ph.D. in Economics from M.I.T. He has published over 100 scholarly articles and book chapters, primarily on the topics of metropolitan housing

xiv

Contributors

markets, racial discrimination and segregation, neighbourhood dynamics, residential reinvestment, community lending and insurance patterns, and urban poverty. He is the co-author of The Maze of Urban Housing Markets 1991, and the forthcoming Why NOT in My Back Yard: Neighbourhood Impacts of Assisted Housing.

Kenneth Gibb ([email protected]) Reader in Housing Economics in the Department of Urban Studies at the University of Glasgow. His research concerns aspects of housing markets and housing policy. In recent years, the main focus has been on housing demand, the economics of social housing and local market analysis. He is the co-author of Housing Finance in the UK: An Introduction (Macmillan 1999) and is currently working on a book entitled The Economics of Housing Policy for Edward Elgar. Kenneth is on the board of the European Real Estate Society and directs undergraduate teaching in his department.

William Grigsby ([email protected]) Emeritus professor in City and Regional Planning at the University of Pennsylvania. Professor Grigsby has had a distinguished career researching housing markets, urban policy and neighbourhood dynamics. He was the author of Housing Markets and Public Policy in 1963 and co-authored The Dynamics of Neighbourhood Change and Decline (1987) with Baratz, Galster and Maclennan.

Joseph Gyourko ([email protected]) Martin Bucksbaum Professor of Real Estate and Finance at The Wharton School of the University of Pennsylvania. He also serves as Director of the Zell/Lurie Real Estate Centre at Wharton and Chair of the Real Estate Department. Professor Gyourko received his BA from Duke University and a Ph.D. in economics from the University of Chicago. His research interests include real estate ﬁnance, local public ﬁnance, and urban economics. Formerly coeditor of Real Estate Economics, Professor Gyourko serves on a number of editorial boards. He is a Trustee Fellow of the Urban Land Institute (ULI), and serves on the Real Estate Roundtable’s Research Committee. Professor Gyourko also serves on the board of EII Realty Securities, a real estate mutual fund, and consults on real estate valuation and securities issues.

Stephen Malpezzi ([email protected]) Associate Professor, and Wangard Faculty Scholar, in the University of

Contributors

xv

Wisconsin-Madison’s Department of Real Estate and Urban Land Economics (http://www.bus.wisc.edu/realestate). Dr. Malpezzi’s research includes work on economic development, the measurement and determinants of real estate prices, housing demand, and on the effects of economic policies on real estate markets. He is co-author with Richard Green of A Primer on U.S. Housing Markets and Policy, published by the American Real Estate and Urban Economics Association.

Geoffrey Meen ([email protected]) Professor of Applied Economics in the Department of Economics at The University of Reading and specialises in the use of quantitative techniques in national, regional and urban housing market analysis. He is the author of Modelling Spatial Housing Markets: Theory, Analysis and Policy (Kluwer Academic Publishers 2001) as well as contributing to the major academic journals in the ﬁeld. Geoff is Head of the Department of Economics at Reading University and also Research Director for the University’s School of Business.

Tony O’Sullivan ([email protected]) Tony O’Sullivan (Ph.D. Economics, Sussex) is Head of Planning for Communities Scotland, a national government agency in Scotland involved in housing investment and community regeneration. He joined the agency in 1989 after holding a lectureship in housing economics at the University of Glasgow for much of the 1980s. He has managed the agency’s research activities, including national house condition surveys. He is co-author of Local Housing System Analysis: A Best Practice Guide (1998, Scottish Homes, Edinburgh) and has researched in the areas of housing subsidies, fuel poverty and the sale of public housing stock.

John M. Quigley ([email protected]) I. Donald Terner Distinguished Professor, and Professor of Economics, at the University of California, Berkeley. His recent research is on the integration of real estate, mortgage, and ﬁnancial markets, urban labour markets, and public ﬁnance. His most recent book, Homeless in California (2001, with Raphael & Smolensky, Public Policy of California, San Francisco) evaluates policies to subsidise housing consumption for households with very low incomes.

xvi

Contributors

Mark Stephens ([email protected]) Senior Lecturer in the Department of Urban Studies, University of Glasgow. He specialises in comparative housing policy and in the impact of European Economic and Monetary Union on housing systems. He is co-author of Housing Policy in Britain and Europe (UCL Press 1995), co-editor of European Integration and Housing Policy (Routledge 1998) and co-author of Social Market or Safety Net? British Social Rented Housing in a European Context (Policy Press 2002). He is editor of the European Journal of Housing Policy and a member of the Co-ordination Committee of the European Network for Housing Research.

Yong Tu ([email protected]) Assistant Professor in Housing Economics in the Department of Real Estate, School of Design and Environment, at National University of Singapore. Her research concerns aspects of housing markets and housing policy. In recent years, the main focus has been on modelling urban housing demand, price discovery, and Asian housing market analyses.

Christine Whitehead ([email protected]) Professor in Housing in the Department of Economics, London School of Economics and Director of the Cambridge Centre for Housing and Planning Research, University of Cambridge. Her work in the housing ﬁeld has ranged from an econometric model of the UK housing system through acting as a member of the government’s Housing Policy Review research team. Lately, her research has concentrated on aspects of housing ﬁnance and social housing provision as well as on privatisation and the relationship between affordable housing and land use planning. Her latest publications include ‘Restructuring Housing Systems’ (York Publishing 2000), which she co-edited with Sarah Monk. She is adviser to the House of Commons Select Committee on Transport, Local Government and Regional Affairs and was awarded the OBE for services to housing.

Nick Williams ([email protected]) Formerly Senior Lecturer in Geography at Aberdeen University, and now a self-employed housing and environmental consultant. His major areas of interest are social rented housing, urban sustainability and environmental aspects of housing. He is co-author of the Sustainable Housing Design Guide for Scotland, (The Stationery Ofﬁce Books, London ), and Greening the Built Environment published by Earthscan, London. Nick is a committee member

Contributors

xvii

and former chair of Castlehill Housing Association, and is also Policy Convenor of the Scottish Federation of Housing Associations.

Gavin A. Wood ([email protected]) A senior lecturer in economics at Murdoch University, Australia and deputy director of the Australian Housing and Urban Research Institute’s Western Australia centre. His main research area is housing economics. The main focus of his recent work has been the design of microsimulation models of housing markets, agency problems in rental housing and the impact of taxation on housing investment. He has published widely in international academic journals, and is on the International Editorial Advisory Committee of Urban Studies.

Preface This volume honours an urban economics scholar who has bridged longstanding gaps between a number of often quite dissimilar worlds – neoclassical and institutional economics, academia and government, Britain and its neighbouring continents, and the two sides of Hadrian’s Wall. By creatively using the knowledge and perspectives of one world to advance discourse in another, Duncan Maclennan has made a singular contribution both to our understanding of urban phenomena and to the design of urban and housing policy. His Housing Economics (1982) changed the way in which we think about urban housing markets. As important as Maclennan’s own research and practice have been to the urban economics and housing policy arenas, his research leadership has perhaps been even more inﬂuential. In directing Britain’s foremost housing research programmes, he developed and implemented a research agenda that has helped to reshape the way in which housing policy questions in Britain are being explored. In carrying out this agenda and through work for OECD, he has brought together housing analysts from different backgrounds and often conﬂicting ideologies to search for common themes and principles that could be broadly applied. Maclennan’s success in bridging intellectual chasms and crossing the boundaries of academic disciplines has been achieved in part through his wit and warmth but also through the contributions of urban analysts whom he either groomed or inﬂuenced. A few of the many persons who have enjoyed and beneﬁted from the collegial settings he has created over the years have chosen to author this book of essays in his honour. These essays do not have a common theme except to the extent that together they reﬂect the range of Maclennan’s interests. Their preparation is also a reﬂection of the affection and esteem in which the authors hold the person they have chosen to recognise in this fashion. William Grigsby University of Pennsylvania Philadelphia, May 2002

Acknowledgements We would like to record our gratitude to a number of people and organisations that helped in the preparation of this book. First drafts of the chapters in this book were presented at a seminar held at the University of Glasgow in November 2001. Communities Scotland provided ﬁnancial support for this event, which signiﬁcantly contributed to the quality of the ﬁnal product. Elizabeth Nicholson gave sterling and stoic support in the organisation and running of this seminar. Gillian Blacklaw provided able secretarial support to the project, and Laure Paterson worked tirelessly on prooﬁng and referencing matters. We received consistent encouragement and support from our commissioning editor at Blackwell’s, Madeleine Metcalfe, and constant forbearance and understanding from our families during preparation that was far beyond the limits of reasonable expectation. As well as providing the substance of this volume, all those who contributed did so with great enthusiasm, and agreed to waive any royalty rights in favour of giving any earnings from sales of the book to Shelter Scotland. We thank them for this generosity. Tony O’Sullivan Kenneth Gibb May 2002

1 Introduction Tony O’Sullivan & Kenneth Gibb

From its very beginnings as a discipline for the study of society, economists have sought to apply economics to real world issues. While the roots of housing economics can be traced back to the work of the original classical and neoclassical pioneers of economic theory, it did not emerge as a separate and recognised area in its own right until very late in the day. After World War II economists, particularly in Europe, turned their attentions to a number of pressing issues of social policy, leading to the emergence of bodies of work on the economics of poverty, transport, education, and health. Applying the insights of welfare economics and emerging theoretical and practical developments, this work sought to address various efﬁciency and equity issues as they related to the everyday lives of millions of people as citizens and consumers. Not in the vanguard of this work, interest in housing as a proper area for study in the UK began to develop in a serious way from the 1970s onwards. Duncan Maclennan has from the ﬁrst been prominent in contributing to and promoting this. The idea behind this book was a wish on the part of many contributors to the ﬁeld to recognise his contributions.

Perspectives on theory Housing is a difﬁcult thing to theorise about. An inherently complex commodity, with spatial ﬁxity a deﬁning characteristic, and asset, investment and consumption dimensions to account for, the economics of housing remains a challenge to those who seek equity and efﬁciency improvements for society to this day. It is a ﬁeld of inquiry where, in Joan Robinson’s terms, thinking in ‘theory time’ is a poor substitute for recognising the role of real, non-abstract, historical time. Moreover, abstracting from the spatial aspects of housing leads, to quote Maclennan, to a curious form of ‘pointless economics’. Good theory necessarily involves jettisoning many of the

2

Housing Economics and Public Policy

assumptions usually made in neoclassical economics, rendering the work more difﬁcult and ‘messy’. Early contributions from Maclennan drew on the insights of the Austrian school, and the work of such authors as Hayeck and Von Mises. The insights produced by the Austrians concerning information, knowledge, ignorance and time are central to his work on housing search and choice (Maclennan 1982). For Maclennan, information problems are central features of the housing market, as are spatial externalities, and transactions costs associated with mobility. In such a world, neo-institutional economics (Oliver Williamson 1975, 1985), evolutionary economics and disequilibrium analysis are often the natural modes of analysis to adopt in order better to understand housing phenomena. Demand for ‘housing’ in the real world necessarily embodies demand for place, for access to public and private services, and reﬂects the fact that housing is in part at least a ‘positional good’ (Hirsch 1976). Maclennan built these factors into his work from the outset, and recognised the implications for the spatial structure of housing markets. Consequently, space and place feature prominently throughout Maclennan’s work, both as a key friction for understanding the complexity of housing and its impact on household and market behaviour. Maclennan has investigated the implications of space at several different housing market scales. An enduring concern has been examining the process of neighbourhood change, within which neighbourhoods are deﬁned as relatively open residential production and consumption spaces. Maclennan has sought throughout his career to apply useful economic models to the problem of interest. One such example at the neighbourhood scale has been his attempt to capture house price change in unimproved properties after nearby properties have received grant aid to improve their fabric, thus measuring the neighbourhood externality (Maclennan 1993b). More generally, in collaboration with American colleagues (Grigsby et al. 1987), he helped develop a comprehensive framework for deciphering neighbourhood change. At the urban or metropolitan scale, Maclennan has both examined spatial aspects explicitly through the early exploration of the concept of housing sub-markets (and the associated concept of ‘product groups’) and used the urban level of analysis as the implicit or explicit focus for microeconomic models of tenure choice, search and housing choice (often in collaboration with Gavin Wood) (Wood & Maclennan 1982; Maclennan & Wood 1982a). In his 1982 text, Maclennan went a long way to establish a richer conception of housing choice, rooted in real time decision-making, search and bidding

Introduction

3

behaviour, the active role of institutions and a more appropriate notion of equilibrium initially developed by Frank Hahn, where market equilibrium is based on learning of information by economic actors. Information, time and space matter as much for supply and its appropriate conceptualisation as they do for demand. Again, Maclennan made it clear from the outset that the reductionist approach of neoclassical theory was particularly inappropriate to many of the tasks at hand. He showed that noise in the market, and uncertainty about the nature and meaning of market ‘signals’ result in endemic inefﬁciency and market failure (Maclennan 1982). These insights remain to be built upon. Deﬁning the product ‘housing’ is a basic issue in the conceptualisation of housing demand, supply, and the housing ‘market’. Equally fundamental is getting an adequate handle on the concept of the ‘price’ of housing. Over time hedonic analysis has become the approach of choice for most economists, but it has some quite fundamental limitations. These certainly do not invalidate the use of this technique, but they do indicate the need to be careful in its application. Maclennan, who has used the technique frequently in his work, provided an early and cogent summary of these limitations (Maclennan 1977b). A speciﬁc issue relating to the price of housing is that of housing subsidies – their deﬁnition, measurement and impacts (not to mention their rationale). Often these take the form of tax concessions. Considerable work in this area has occurred in the last 20 years, following early critique of the efﬁciency and equity effects of such subsidies (in the UK context see for example Robinson 1981; Maclennan & Wood 1982b; Maclennan & O’Sullivan 1987; O’Sullivan 1984, 1987). Much subsidisation of housing, however, particularly in a European context, has been tied into direct provision through public or social housing. This has raised issues not just about the welfare implications of the pricing of housing, but the comparative efﬁciency of public versus private provision. On pricing per se a considerable literature has developed, to which Maclennan has contributed (Atkinson & King 1980; Grey et al. 1981; Ermisch 1984; Hills 1991, 2000; Maclennan 1986b). On the question of public versus private provision, traditionally non-market providers of housing have operated through bureaucratic hierarchies, providing a rich opportunity to apply Coasian transactions costs concepts and insights from the wider work of Oliver Williamson to the speciﬁcs of organisational efﬁciency in public housing. Maclennan contributed to such analysis in the late 1980s in considering the comparative efﬁciency of different housing providers in the UK (Centre for Housing Research 1989).

4

Housing Economics and Public Policy

While the microeconomics of housing has developed apace over the last two decades, attention has also been given to modelling at the regional and macro scale. At the national scale of analysis, housing in the UK (as elsewhere) was transformed in the 1980s by a process of ﬁnancial deregulation. Collaborative work by Maclennan with Geoff Meen, John Muellbauer, Mark Stephens and others identiﬁed many of the connections between housing and the macro-economy and helped forge a bridge between housing specialists and those macroeconomists coming to the housing sector for the ﬁrst time. This also helped policy makers who needed to understand the micro-foundations of mortgage market change, housing equity withdrawal, house price volatility and their transmission mechanisms into consumption, indebtedness and the labour market (Maclennan et al. 1997).

Policy and practice Much of the interest in housing economics that has developed over the last two decades has emerged on the back of considerable intervention in housing by central and municipal government. Policy makers of all political hues, especially but not exclusively in Europe, have seen it as desirable if not essential to intervene in, regulate, suspend, replace or compete with various aspects of the housing market. Maclennan has been particularly active in seeking to inﬂuence policy, especially, but not exclusively, in a UK context. In doing this, being based at the University of Glasgow for most of his professional working life is not the only link between Maclennan and the classical economists. In his conception of policy there is much to link Maclennan’s approach to the classical school. From the ﬁrst, any failure to distinguish technical from ideological positions in policy analysis has been rejected as at best sloppy intellectualism and at worst sharp practice. This does not imply however that he has promoted a narrow ‘technical’ interpretation of the economist’s role in policy formulation; on the contrary he has been a constant advocate of a broader ‘political economy’ interpretation of the role. Moreover, the use of assumptions regarding the operation of housing (and wider economic) systems to progress an argument, where the basis should be the results of applied research, has usually been given short shrift. In this regard, Maclennan has been very aware of the relatively poor base provided for applied research by ofﬁcial statistics and has indeed been a stern critic of government. In 2001 Maclennan was involved in redesigning the Right to Buy (RTB) in Scotland as an element of what was to become the Housing Act (Scotland)

Introduction

5

2001. Starting from an examination of the available evidence he commented: ‘The RTB has been a source of enduring controversy over the two decades since its introduction. Views on policy proposals can diverge for at least three different reasons. First, individuals or organisations may have different ethical or political views about an issue and facts or evidence may have little relevance in resolving such differences. Secondly, many policy decisions involve changes with complex, long-term effects on the economy and society. Individuals with the same values may reach different conclusions because their model or analysis of impacts differs from each other. Again this requires judgement in choices, but it also requires the participants in a debate to be clear and consistent in their analyses. Finally, even with the same ethical values and analytical approaches, individuals may reach different conclusions if they have different facts or evidence at hand.’ (Maclennan et al. 2000, p. 1) These views echo those expressed in his 1982 text (which, coincidentally, used the then very new RTB issue to illustrate the point). Maclennan has consistently rebuked both economists and critics from other disciplines who have sought to interpret or represent the role of the economist in policy discourse as being focused on competitive housing markets, the use of deductive partial or general equilibrium models, or to be unthinkingly centred on Pareto-based social welfare functions. For him, ‘… it is not enough to restrict the role of the political economist to accurately describing market behaviour for policy makers – although such action may be an improvement on proffering advice largely based on deductive reasoning and assumption. Instead, applied economic analysis of policy (or political economy) requires the analysis of a more detailed set of questions. For instance the analyst must enquire as to what are the objectives of government, what role housing policy plays in attaining such objectives, who actually makes housing policy, what economic and political constraints restrict action, what advice and information exists and what model of the housing system is used by policy makers? These questions are not to be viewed as minor concerns for applied economists. Economists may deduce conditions for optimality or second best conditions for the design of housing policies ad inﬁnitum but unless they understand how economic analysis and research is absorbed and transformed by the policy

6

Housing Economics and Public Policy

process they run the risk of having their work, at best, ignored or, at worst, misused.’ (Maclennan 1982, p. 142) Policy inﬂuence, then, must start from an understanding of the institutional, administrative and legal contexts from which policies emerge. Ethical positions on policy issues are acceptable (indeed unavoidable) but should be explicit and honest. Value judgement should not be dressed up as technical issue, and ignorance on empirical matters of importance must be acknowledged and must constitute the soil to be tilled by the applied researcher. This view of the policy role is demanding above all else of honesty, and for this reason is brutally demanding in application. While Maclennan has been clear from the start of his career that governments can quite legitimately hold to non-Paretian objectives, he has also been steadfast that these objectives must be capable of clear operational interpretation. This in Maclennan’s eyes is perhaps the key part of the bargain policy makers have to deliver, for the policy-making process to have any hope of working to the beneﬁt of the citizen – the impact on whom always remains the focus of Maclennan’s work. In his writings but even more so in direct dealings with senior government administrators, ministers and politicians, he has delivered this message in an uncompromising way. Given this, other issues can be largely worked around – recognising there is plenty to be so worked. For a start, the structure of government and the distribution of policy responsibility across different tiers and geographies of government make the achievement of coherent policy intrinsically difﬁcult. Joined-up government is a popular current slogan, but also a desideratum recommended by Maclennan over 20 years ago – with due recognition of the inertia built into the existing system, and the consequent need for an institutional economics perspective on the problem. Writing in the UK policy context he noted: ‘[The] complex organisational model for housing policy in Britain would not necessarily be inadequate if some centrally determined operational principles for policy existed such as market pricing or income/wealth related subsidy systems. But where policy, as in Britain, lacks a central organising principle then the distribution of subsidy reﬂects neither house type nor household characteristics but is an outcome of chance, history and in particular of tenure selected. Thus, whilst it has been argued that housing market intervention may be rational and required, it is difﬁcult to believe that the present system could not be improved or it may even

Introduction

7

be that the imperfect market system that the policy is designed to circumvent would perform more effectively as an agent of allocation and redistribution. However it is probably not possible to answer this question in the UK because, as in most ﬁelds of policy in the country, there is often more concern with the level than the quality of public expenditure.’ (Maclennan 1982, p. 167) In his 1982 text, he categorised housing policy into a number of different types of successively greater generality. These were: • Removing speciﬁc market failures through taxation, subsidisation, or regulatory action • Replacing fundamentally unﬁxable market with administered allocation mechanisms • Using housing as primary vehicle for achieving non-allocational ends, such as income redistribution, or regional growth. His subsequent work has demonstrated that simple or pat conclusions on the relative attractiveness of state versus market-based policy solutions, while endemic in policy debate, invariably miss the point. In local markets, for example, it is not clear that public monopoly provision would be in any way superior to a private sector version. In a sense the policy question of interest to Maclennan has always been not whether housing (and regional economic) outcomes can be optimised by state intervention or forbearance, but whether, working in a real world context, market failure or state failure constitute the least bad means to achieve clearly articulated ends. To answer this question, he made an early plea for what has come to be widely referred to as ‘evidence based policy’. Starting in his 1982 text, Maclennan made a point that has continually recurred in his work: that there is a need for a co-ordinated provision of market and non-market information to drive market allocation decisions, and to inform public action. The absence of this information base resourced (if not necessarily discharged) as a function of government has, he has argued, led to poor investment planning in the private as well as the public sector. In a public policy context, the consequences are as follows: ‘In this context for policy formulation, there are two alternatives at least for formulating strategies. First, ministers … will listen to the arguments of pressure groups and interested parties. Power, in this case, not to the

8

Housing Economics and Public Policy

people, but to the groups with organisation and capital to present the most convincing (not the best) case. Alternatively, ministers with an inner circle of policy advisors may form a view as to what is appropriate policy. These views may be based either simply on ideologies or on a limited or even partially researched view of the housing system.’ (Maclennan 1982, pp. 168-9) Part of the problem as he saw it was the absence of appropriate training for those engaged in policy making and implementation, from the strategic to matters of detailed operation. He certainly demonstrated the personal courage of this conviction, in one instance telling an audience of around 500 not-for-proﬁt housing practitioners that in his opinion they couldn’t strategically plan their way out of a paper bag! More constructively, his view has been expressed as follows: ‘It must be of some concern within central government that many ofﬁcials responsible for policy have no initial familiarity with housing and that, within the British Civil Service, transfer often follows mature acquisition of insight. Similar comments … apply within local authorities and the lack of formal training within public housing management is now the cause of some concern …’ (Maclennan 1982, p. 168) This in part has led to the establishment of major programmes of housing training at Glasgow University from the early 1980s onwards, integrating theory, policy and practice. These covered both the needs of housing management and development practitioners through to the tutelage of civil servants and local authority ofﬁcials at a postgraduate level. Maclennan was heavily critical of policy reﬂected in key pieces of housing legislation in the late 1970s and early 1980s (Maclennan 1982). He saw it as a professional responsibility on him and fellow economists to present government with usable models, methods and material that can improve the quality of public policy. His subsequent career has been geared to the dispensation of this responsibility at the national and international level. His success in this regard is perhaps best reﬂected in the pivotal design role he played in the development of the Housing Scotland Act 2001, which gave him the opportunity, albeit in a purely Scottish context, to inﬂuence the structure of the Right to Buy, of which he was so critical in 1982.

Introduction

9

Contributions to this book The contributors to this volume have all had the opportunity and pleasure of working with Duncan Maclennan at some time in their careers. Each of the chapters picks up on an area of interest to him over the last 20 years. Collectively, these chapters cover much of the landscape of housing economics and housing policy.

Urban housing models The chapter by Kenneth Gibb assesses the different types of urban housing models that have been developed. Gibb considers the role the ‘standard’ access-space approach has played as the starting point for analysis of urban housing in much of the literature. The chapter makes a case for more extensive use of ﬁltering research in the future investigation of urban housing systems. It also looks at the latest generation of urban simulation models, in particular those associated with Anas and Arnott, and presents a simpliﬁed metropolitan model of Greater Glasgow developed by Gibb, Meen & Mackay (2000). More fundamentally however, Gibb argues that the evolution of these different types of model has failed to address many of the concerns voiced by Maclennan (and subsequently reiterated e.g. Maclennan & Gibb 1994). It is not that these models fail to give useful insights, but that by and large their strengths and weaknesses remain as proﬁled two decades ago. While ﬁltering models perhaps as noted have a greater potential for providing insights than they have so far delivered, the whole ﬁeld would beneﬁt from new thinking. The future development of spatial models of the housing market will need to incorporate adjustment and disequilibrium on both sides of the market, information imperfections, reﬂect the importance of real time and process in market evolution, and capture the key role of housing market institutions in a way that can be sensibly operationalised. This is a formidable challenge.

Segmentation, adjustment and disequilibrium Urban housing markets are generally thought of as segmented in one or more dimensions. This can be in product group terms, or in terms of quality hierarchies, where the key distinction is made in order to classify the market into sub-markets involving dwellings that are more or less substitutable as far as consumers are concerned. In her chapter, Yong Tu provides an analysis of the causes and consequences of housing market segmentation. Using the well-

10

Housing Economics and Public Policy

known conceptualisation of housing as a composite commodity, she deﬁnes sub-markets in terms of underlying attribute prices. A given attribute has a constant hedonic price within a sub-market but across sub-markets it will vary. One important implication of the analysis of segmentation Yong Tu provides is that disequilibrium is likely to be the norm. She examines the properties of such disequilibrium in the housing market. Turning to the thorny problem of identifying housing sub-markets in practice, Yong Tu evaluates the main alternative methods and concludes that there is nothing that intrinsically distinguishes one as superior to another. Of more signiﬁcance, her analysis suggests that urban house price indices should be constructed at sub-market level, and is again strongly persuasive that disequilibrium should be explicitly factored into analysis of urban housing. Yong Tu also argues that a disaggregated framework is the proper basis for analysing the local neighbourhood impacts of urban and housing policies.

Transactions costs John Quigley’s chapter deals with the question of transactions costs and their signiﬁcance for the housing market. Apart from the costs of search, consumers must deal with legal negotiations and bargain with other parties. In addition, because choice is infrequent and information or knowledge of the housing market decays, costs linked to uncertainty and expectations are important (if inherently difﬁcult to measure). His chapter sets out a taxonomy of these costs and reviews the empirical literature on their individual magnitudes. He reports, for example, that the average costs of negotiating and legal transactions in the US housing market vary from 6% to 12 % of house value. However, these costs may be relatively small in comparison to moving and psychic (adjustment) costs. Quigley presents a simple model based on consumers making move/stay decisions by comparing the discounted utility gains forgone by not moving with the discounted transactions costs that a move will incur. This, he suggests, gives an alternative basis for measuring the magnitude of transaction costs facing households to deriving this information through survey work. Quigley also considers whether transactions costs in the housing market involve signiﬁcant costs for the wider economy, and whether policies can be developed to lessen any such costs. Although Quigley acknowledges some limited theoretical support for Oswald’s contention that reduced owneroccupier mobility is the source of higher associated unemployment among

Introduction

11

homeowners in different OECD countries (Oswald 1997b, 1999) (where debt constraints, such as negative equity, effectively act as arbitrarily high transactions costs inhibiting mobility), he ﬁnds no credible empirical evidence that home ownership causes high unemployment. Quigley argues that government should act to reduce transactions costs in the housing market, however, facilitating ﬂows of information, and in the development of standards that can make transactions simpler.

Hedonic pricing models Stephen Malpezzi reviews research on one of the most important tools in housing economics: hedonic price models. There is a voluminous literature concerned with house price index construction and a smaller ﬁeld of work examining the underlying demand for individual attributes. In his chapter, Malpezzi provides a clear exposition of the fundamental conceptual and empirical issues often ignored when undertaking applied work in this area. In particular, he considers the functional form problem (there is no obvious a priori functional form to adopt and the one chosen is usually on the basis of convenience alone). Malpezzi reviews the econometric issues associated with identiﬁcation, mis-speciﬁcation, non-robustness of co-efﬁcients, nonlinearity and disequilibrium tendencies. Malpezzi provides an overview of the main applications of hedonic analysis in housing and identiﬁes areas that would repay further study. He draws on Maclennan’s (1977b) work to suggest, given the continuing dependence of housing economics on this class of model, that hedonic models should de minimis possess three features: their use in any speciﬁc application should have a ﬁrm theoretical basis; their speciﬁcation should be complete and if not the practical consequences should be fully explored; and, perhaps most critically, the design of a pricing model should be ﬁt for its intended purpose. These recommendations are pertinent to a literature that is too often consumed by a sterile battle between protagonists of different index schools, the performance of which are not always strictly comparable.

Housing in the macroeconomy In his chapter, Geoff Meen surveys the relationship between housing markets and the macro-economy. Meen organises his review around ﬁve themes: the relationship between house prices and consumer expenditure; the effect of housing on the labour market nationally and regionally; national differences in housing markets and their consequences for EMU; the contribution

12

Housing Economics and Public Policy

of housing to economic cycles; and possible links between housing and the spatial concentration of business start-ups. He calls into question the feasibility of house price prediction, house price being the key variable linking housing to the macroeconomy, and moves on to investigate the implications of unpredictability and to suggest a different basis for housing market analysis starting from the bottom up, recognising that local markets cannot be treated as if they are all effectively the same. Meen examines the scope for using complexity theory as a basis for simulating local housing markets, characterised by heterogeneous agents, non-linear outcomes (such as ‘tipping’) and multiple equilibria (including prolonged disequilibria). Meen argues that tipping and thresholds are important processes underpinning neighbourhood housing market change from decline through normalcy to gentriﬁcation. While arguing that the reductionist neo-classical models may well be inappropriate for many aspects of housing economics research, however, he does point out that they continue to have relevance under speciﬁc circumstances and that ‘a slavish attitude to any approach is rarely likely to be optimal’.

Taxation, subsidies and housing markets Gavin Wood examines taxation as one of a set of frictions that impact on the efﬁciency of housing market outcomes, as well as having important distributional consequences. His chapter demonstrates the regressive and often counter-intuitive effects of housing taxation. He then uses tax arbitrage models to analyse tenure choice. His literature review and subsequent analysis of the distributional incidence of taxation on housing shows that, while the user cost effects of housing taxation may be well known, the systematic variations in landlord user costs associated with differential tax treatment are less well known or studied despite their importance for rental investment. In the discussion of tax-induced efﬁciency distortions, Wood makes a case for rental housing supply segmentation arising from ‘clientèle effects’. Furthermore, there are ‘lock-in’ effects associated with this segmentation because capital gains taxation is rarely levied on an accruals basis. Wood goes on to review the literature on the size of the welfare loss associated with taxation expenditures for home ownership, assessing the proposition that tax-induced housing investment (often in second-hand housing) may have crowded out industrial or ‘productive’ capital. He ﬁnds the evidence to be conﬂicting, suggesting that housing investment generates both positive and negative externalities in the wider economy.

Introduction

13

The economics of social housing In her chapter, Christine Whitehead applies economic principles to another question of considerable policy importance: what is the rationale for social housing? Whitehead concentrates on three issues: the efﬁciency rationale for government intervention, the role of housing as a method of redistribution, and the insights that economic theory can provide into appropriate governance structures for housing provision and management. Whitehead shows that housing is subject to a signiﬁcant number of market failures, and that their impact is, on the balance of evidence available, probably large enough to matter; but that it is far less obvious that social investment in housing will necessarily solve the problems these failures create. The political case for the provision of social housing has usually been made in terms of the efﬁcacy of housing as a direct means of redistribution. Whitehead shows that the relative merits of income versus supply subsidies for achieving distributional objectives are less obvious in the case of housing than many assume, and that the most appropriate way to achieve distributional objectives comes down to empirical answers to questions regarding the relative importance of market failures in the context of speciﬁc economies and housing markets. Moreover, the possibility of state failure must also be factored into any analysis of the appropriate response to market failure, or assessment of the form of intervention needed to achieve an agreed distributional objective. What is required in a housing context is a proper evaluation of governance issues. As Whitehead notes, in reality, any governance system is imperfect. It is therefore a matter of analysing the different attributes of market, regulatory and administrative systems in particular social, political and market contexts to determine the least-cost approach. Whitehead concludes in the UK context at least that in the future there should be less emphasis on social housing per se and more on managing markets, ensuring adequate investment and enabling different types of providers (public, non-proﬁt and private) to compete on cost and quality.

Neighbourhood dynamics In his chapter, George Galster considers the notion of ‘neighbourhood’ from an economic perspective. Deﬁning neighbourhoods as bundles of spatially-based attributes (including such things as class proﬁle, tax/public service levels, demographic proﬁle and the political activity characteristics of residents as well as property-based characteristics), Galster shows that

14

Housing Economics and Public Policy

neighbourhoods have a number of important features. These are that attributes vary in their durability and ability to be priced, that they are valued in a relativistic (cross-neighbourhood) way by consumers and that the spatially-based attributes comprising neighbourhood can change by the very act of consuming them. Galster shows the implication of these features to be that the costs and beneﬁts associated with consuming a neighbourhood largely originate externally to the consumer in question. Galster presents neighbourhoods as dynamic and subject to constant change, and at the same time highly risky for potential consumers and investors. He conceives urban neighbourhoods as being situated within a broader segmented housing market, with change at the neighbourhood level driven fundamentally by external shocks transmitted across the market. Once initiated by such a shock, within-neighbourhood micro processes that can be represented through complexity models embodying non-linear threshold effects ensue. Galster argues the intrinsic nature of neighbourhoods is such that they are not likely to produce socially efﬁcient outcomes. Moreover, because neighbourhoods are valued in a relativistic way, poorly conceived policy might have zero sum effects. On the other hand he points out careful targeting can be highly effective in assisting neighbourhoods on the cusp of catastrophic decline.

Home ownership and constraints to tenure choice Joe Gyourko looks at the determinants of access to owner occupation in the US. Until quite recently, permanent income and the relative cost of owning versus renting were thought to be the dominant factors. Gyourko reviews recent theoretical and empirical work that shows it is net worth and asset price that drives the transition from renting to owning. This view of the world implies that attempts to improve access to owner occupation through policies aimed at reducing payment to income ratios are misconceived. Leaving aside the question of whether such an objective is in fact appropriate, Gyourko shows that reducing the required down-payment on purchase should be considered the more effective policy instrument. Gyourko goes on to consider the implications of this new view for explaining racial differences in US home ownership rates, and ﬁnds that these can be eliminated once household wealth is properly controlled for. Properly controlling for wealth in this context means allowing for racial differences in intergenerational wealth transfers between parents and children, as well as accounting for a household’s own net worth. Because of these differences in wealth, black households are far less likely than their white counterparts

Introduction

15

to apply for a mortgage, and Gyourko reports evidence that shows this is far more signiﬁcant than discrimination in the mortgage market in explaining ownership differentials. Gyourko also considers trends in the affordability of ownership in the US since the 1970s. Rather positive conclusions on this can be derived from a prominent national interest rate-based index measuring housing costs to the median income household purchasing the median home. In contrast, he presents evidence showing that affordability conditions have varied across the income distribution, with low-income households having fared badly. This ﬁnding reinforces the signiﬁcance of Gyourko’s conclusions regarding the relevant budget constraint facing households who are seeking to become owner-occupiers.

Planning regulation and housing supply In his chapter, Glen Bramley discusses housing supply. While supply is clearly of central importance to housing economics and the operation of housing markets (as well as more generally), Bramley points out that it typically receives much less attention than demand, and suggests this might derive from the difﬁculty of applying deductive modelling to this subject. Bramley reviews past economic research on new housing supply, and discusses the operation of land use regulation in Britain, which he contrasts with systems in other countries. While land use regulation is likely to affect the supply of land for new housing development, land and house prices and housing densities, the scale of these effects and their welfare interpretation is not certain. Bramley concentrates on the measurement issues and employs a panel model to analyse the impact of the amount and type of available land on new housing supply and house prices in England. The data are organised spatially on the basis of 90 ‘zones’ constructed as simple aggregations of local authority districts and Bramley builds a straightforward supply and demand system, on the basis of which he concludes the price elasticity of new build supply is low in Britain, and may be getting lower. Bramley also ﬁnds low land price elasticities and draws a number of conclusions of policy signiﬁcance from his model. These include that land availability through planning is a strong driver of development, that greater land supply reduces house prices to a modest extent only, and that emphasis through the planning system on residential development on urban and recycled land carries a signiﬁcant output penalty.

16

Housing Economics and Public Policy

Housing planning In his chapter, Tony O’Sullivan looks at local housing planning in the UK. There has long been a presumption in the UK that such planning is necessary (and the chapter by Galster supports this conclusion), and that municipal government is the appropriate agency to carry it out. Having considered the case for such planning, O’Sullivan looks at what it is expected to achieve, and compares this with what it has actually delivered. Housing planning and the related land use planning that occur in the UK are shown to be almost entirely devoid of economic content. O’Sullivan details the areas where it is reasonable to expect economics to play a role in housing planning as conceived in the UK and contrasts this with the economic tools actually available. However, he points out that even with better tools a major gap between requirements and reality might be expected to remain unless incentives structures between and within levels of government are adjusted. He concludes that housing planning per se is needed, and that the American apathy towards such planning is not a better approach. However, the emphasis should be on the process of planning, and plans produced should be seen as contingent on unknown and currently unknowable future events.

The Right to Buy The ﬁnal three chapters in this book look explicitly at different aspects of the relationship between research and academics on the one hand, and government and policy development on the other. The chapter by Williams looks at the Right to Buy in the UK, a policy under which local authorities have been required to sell rental stock at a substantial discount to market value. Right to Buy has been the most signiﬁcant and contentious housing policy in the UK in the last 20 years. Williams shows that between the introduction of the Right to Buy in 1980 and the end of the century there was little connection between research and policy development. It was introduced to increase owner occupation, pure and simple, not to address any speciﬁc equity or efﬁciency issues in the UK housing system. Key restrictions on sale under Right to Buy were adopted at the outset on the basis of skilful political lobbying, not hard evidence of a problem to be addressed. Meanwhile, subsequent evidence on the negative and unintended effects of Right to Buy was largely ignored.

Introduction

17

Academic (and housing professional) reaction to Right to Buy was largely hostile, and Williams argues this too was at least in part ideological rather than evidence-based. While much research has now been done, key research into the local effects of Right to Buy, called for as early as 1981, still remains to be undertaken. Williams considers recent developments in Scotland where the establishment of the Scottish Parliament in 1999, with housing as a devolved function, have provided an opportunity for more rational arrangements for the Right to Buy to be established. In 2001 a Housing Act was passed that signiﬁcantly amended the terms under which Right to Buy can be exercised in Scotland. He shows how these amendments were based for the ﬁrst time on serious examination of the available research evidence. Whether or not the revisions to Right to Buy lead to better efﬁciency and equity outcomes in the Scottish context remains to be seen. The more general light Williams’ chapter throws on the relationship between academics and policy makers also raises many interesting questions.

The political economy of housing research Donnison and Stephens address some of these questions in their chapter. They consider the scale and scope of housing research undertaken over time, in the context of changes in the political and academic environments for such research. In doing this, they are able to identify some of the main factors responsible for shaping the evolution of housing research in Britain. They situate housing research clearly in the time-speciﬁc cultural, political and economic circumstances of the countries in which it takes place, but argue that over time academics have faced increasing competition as a source of knowledge for policy makers, and greater difﬁculty in disseminating their ﬁndings to the wider public. Reminding academic colleagues ‘who think of themselves as analysts of housing problems and policies, that they are applied social scientists for whom their own profession is also a proper subject for study’, Donnison and Stephens argue ‘the comparative advantages remaining to academics are their opportunities – and obligations – to gain a broad understanding of the workings of their society, and to speak uncensored truth to power and to the public at large’.

Policy and academia The ﬁnal chapter comes at the relationship between academics and policy makers from a completely different direction. For Best, the issue is one of how to maximise the relevance and impact of research ﬁndings on policy development. He acknowledges the tensions that exist between academics

18

Housing Economics and Public Policy

and policy makers, but sees them as in something of a symbiotic relationship. Sometimes academics take a direct and active part in the development of policy, even to the extent of taking up active political roles. Best outlines the pros and cons of these situations, and considers some of the personal qualities needed to make this transition successfully. He goes on to discuss how research undertaken by those who do not wish to take an active political role feeds into policy development and sees this as being determined by four factors: who pays for the research (and what their motive is); how it is presented; the timing of research ﬁndings; and the persistence with which they are presented to policy makers. Best makes the case that the skills needed to bring research to bear on policy making are not the same as those necessary to do the research in the ﬁrst place, and raises the question of strategic alliances with intermediate agencies that specialise in driving research into policy development. As such agencies are thin on the ground, he suggests an alternative might be to promote the development of specialist units within universities to carry out the task.

Recurrent themes and key issues for the future The contributions to this book cover a very broad range of issues in theory, policy and practice. In terms of the contributions Maclennan has made to housing and urban economics, one can identify a number of common themes that recur across chapters. One relates to the ﬁtness-for-purpose that models of housing markets exhibit. Malpezzi points out that ﬁtness-for-purpose is a fundamental criterion in deciding which method to use in the construction of a house price index. Simple frictionless neoclassical models are neither right nor wrong, and to try and categorise them as such would be to miss the point. For some purposes, as Gibb, Galster, and Quigley show, they are extremely valuable. But for much housing market analysis there is a need to accept that many of the assumptions made in the development of such models actually dispense with the real problems of interest. The chapters by Yong Tu and Quigley, Galster, Meen and Bramley, in different ways, testify to this. A second relates to the issue of state versus market failure, and choice of the ‘least worst’ policy option in any given context. The chapter by Whitehead deals explicitly with this issue, but in essence, this is an issue underlying the chapters by O’Sullivan, Galster and Meen. If neighbourhoods are socially inefﬁcient and there is a concern with the local equity effects of housing consumption and production, what should be done about it? Is there a case for state-led planning, or will the outcome be no better, or even worse than

Introduction

19

that which the market will deliver? In practice of course there is no deﬁnitive answer to this and the issue must be considered on a case-by-case basis – another instance of being aware of ﬁtness-for-purpose. A third relates to the issue of evidence-based policy. The chapters by Wood, Gyourko and Williams show that in the absence of sound evidence policy positions can be irrelevant to the problem they are attempting to address, or generate unintended and unwanted distributional and/or efﬁciency effects. Developing evidence-based policy is often no simple matter, however, as the chapter by Bramley shows. A fourth concerns the more general issues that the chapters by Williams, Donnison and Stephens, and Best throw up in terms of the appropriate relationship between academia and policy makers. There is a constant tension between the world of politics and the world of the academy. Should research be the handmaiden of policy, or the guardian of the public interest? Can it be both? In practice, there must be times when academia speaks truth to power, and times when power returns the compliment. Again, the relationship between policy-making and research must in any speciﬁc situation be one that is ﬁt for the given purpose – but we must all be aware of the rules of the game being played. The individual chapters in this book also show how housing economics has developed over the last two decades. An important complementary question to ask is how it might develop in the years to come. Our crystal ball is as clear or as foggy as the next person’s with an interest in the ﬁeld, but we offer the following thoughts and hopes on the matter. As with any applied area, data and access to it is of fundamental importance to housing economics. In the last 20 years there have been remarkable leaps in computing power and ease of use for even the most demanding econometric and mathematical software. With some important exceptions, this has not always been accompanied by improvements in the quality of housing market data sets. This is partly the result of commercial conﬁdentiality but it is also a consequence of data providers failing to recognise the policy and practical relevance of better information. In Britain, the pursuit of evidencebased policy is becoming reﬂected in more open access to public data, in recognition of the greater need for good empirical information at local levels and in an increasing amount of longitudinal data collection. However, even in 2001, the Census was carried out without any question on household incomes or resources. Our hope for the future is that technological developments in statistical and GIS software will marry with increased access to improved data in the UK – but this is unlikely to happen evenly or smoothly.

20

Housing Economics and Public Policy

We suspect that space will become more rather than less important in both theory and policy terms. While an increasing amount of research in theoretical housing economics is attempting to allow for space in a meaningful way at different levels of aggregation, it is also becoming increasingly clear that housing analysis has to cohere across levels of aggregation (Meen 2001). An integrated framework of analysis from the level of urban neighbourhoods, through sub-markets and segmented housing markets to an adequate representation of regional and ultimately national housing markets is a long way off, but a prize for which more people are reaching. Galster sketches elements of it in his chapter, and Meen is actively working on aspects of this agenda, but as we say, it will be the work of many years and many hands to achieve. One reason why it is likely to come is an increasing awareness, certainly in the UK context, that policies developed to operate at different spatial levels can inadvertently work against, even neutralise one another, or alternatively greatly enhance each other’s chances of succeeding. Maclennan points out that ‘[ignoring] geography has a long tradition in British policy making’ and that ‘[for] most of the period since 1980, until the last few years, economic policy was constructed as if the economy existed on the head of a pin’ (Maclennan 2002, p. 1). But Britain, in common with a number of other advanced European economies, is beginning to accept the idea of ‘territorial management’ as a legitimate, indeed inescapable policy objective. Community regeneration policies (for reasons Galster and Meen elucidate) cannot work without an explicit spatial dimension: ‘There are also major gains to be had from setting community level initiatives in a better-understood geography of wider systems, for instance in a city-regional framework. More often than not a failure to do this means that policies concentrate on local physical actions and facilities rather than convincingly connecting regeneration activities to wider labour markets, transport systems and the environment; but if they are to succeed long term [these are] precisely the wider system connections that have to be made.’ (Maclennan 2002, p. 7) And of course both local and city level policy has to connect in some meaningful way with national and international developments. ‘We need a new framework of policy thinking and development which marries the impact of top down external forces and events with the creative and recursive effects which clearly push from the bottom up.’ (Maclennan 2002, p. 6)

Introduction

21

With a continued growth in understanding by policy makers of the signiﬁcance of space, not at one, but at many levels simultaneously, there will be increasing demands for models and theory to provide an evidence-base for policy work. And this last requirement is the ﬁnal big pointer towards the future. While academic independence must never be taken lightly, the trend to evidencebased policy will gather steam, and hopefully create a positive dynamic between the academic and policy-making worlds. As new benchmarks are set in the evidence-base of speciﬁc policies, greater demands will be placed on research to furnish the needed information. More sophisticated policy positions will increase the demand for more sophisticated and realistic representations of housing markets. The meaning of ‘ﬁtness for purpose’ is shifting in policy terms, and with it the goalposts for what is acceptable as a representation of housing markets will shift also.

2 Urban Housing Models Kenneth Gibb

Introduction The fact that different approaches to modelling housing markets exist probably reﬂects the complexity of the commodity ‘housing’. Olsen (1968 p. ·612) took this complexity as motivation to develop a competitive model consistent with the standard microeconomic paradigm and not restricted by the ‘idiosyncratic concepts used by housing specialists’. Initially, spatial models of the housing market focused on long run equilibrium explanations of the impact of commuting on residential markets for space (Alonso 1964). Other research de-emphasised the spatial aspects of housing in favour of attention to its durability and dynamics, emphasising the effects of ‘ﬁltering’; the process by which a dwelling’s attributes and occupants change over the property’s lifetime (Arnott, Davidson & Pines 1983; Grigsby et al. 1987). The importance of heterogeneity has been reﬂected in work concerned with the creation of price indices, and the study of sub-markets. In these areas, the ‘hedonic’ approach has been of central importance. The purpose of this chapter will be to consider and assess different classes of urban housing model; in particular, neoclassical spatial models and the more heterogeneous ﬁltering tradition. Both literatures have included important contributions from Duncan Maclennan (Maclennan 1982; Grigsby et al. 1987). These different ways of analysing the urban housing market have provided controversy and insight alike; yet they are rarely considered as classes of model or explanation of housing phenomena. The chapter also assesses progress in the development of computer-based simulation models of urban housing markets. Maclennan has constantly emphasised the potential importance of such models and their relevance to housing planning and policy assessment if developed with sound micro

Urban Housing Models

23

foundations capable of reﬂecting market processes (e.g. Maclennan 1986a; Maclennan & Gibb 1994). As they have developed, these models have often embodied elements of either or both of the main classes of model considered in this chapter but additionally they have sought to provide frameworks to test policies, scenarios and the consequences of system shocks. This is an area in which there has undoubtedly been signiﬁcant progress in the last 20 years (Ingram 1979; Anas & Arnott 1991, 1993a, b, c, 1994, 1997).

Urban residential trade-off models There is a considerable literature regarding the monocentric residential model or ‘new urban economics’ (Alonso 1964; Mills 1967, 1972; Muth 1969, 1985; Evans 1973; Solow 1973; Wheaton 1974; Quigley 1979; Fujita 1987; Straszheim 1987; Wheaton & DiPasquale 1996; McDonald 1997; Anas, Arnott & Small 1998) and I do not intend to revisit it extensively. Rather, in this section, I focus on a limited number of central conceptual and empirical issues that arise from this modelling approach. Maclennan (1982) argued that the (access-space) model ‘represents the real starting point for an analysis of local urban housing markets. The model is currently the dominant paradigm of urban economic research in North America and it is thus important to show its uses, merits and demerits’ (p.·7). Appraisal of urban economics and urban real estate texts (e.g. Wheaton & DiPasquale 1996; McDonald 1997) shows that this remains the case 20 years on.

The basic trade-off The essence of the model (following Henderson 1985) is that within a simpliﬁed model of urban land use, utility maximising households offset employment accessibility to a central workplace and retail location (measured in monetary and time commuting costs) against the consumption of residential land (or housing) space. The model generates a long-term spatial structure for household distribution that can be empirically tested. It can be analysed in terms of comparative statics and has been extended in a number of ways to increase the realism of the basic structure (e.g. with polycentric rather than monocentric economic activity locations). In the basic model, the modelled city structure is assumed to be a ﬂat and featureless plain where all employment is located in the ‘Central Business District’ (CBD) and, with the assumption of no congestion, commuting

24

Housing Economics and Public Policy

takes the same unit of time (t) to travel a given distance in any direction. Households in such a world are assumed to maximise utility subject to an income and a time constraint, consuming three generalised goods: the city’s own traded good (x), the city’s imported good (z) and housing services (h). Only the price of the latter varies across urban space. Housing is viewed as a consumption good with a locational amenity. The only amenity choice is leisure consumption, which is a direct function of commuting time to the CBD. In this world, the ‘rental on housing implicitly prices both housing services and access, or leisure consumption’ (Henderson 1985, p.·3). Thus, the unit price of housing p(u), at distance (u) from the CBD, varies spatially as leisure varies and leisure is the time T left after working and commuting. The representative consumer has leisure consumption e(u) at distance u from the CBD equal to: e(u) = T – t(u)

(2.1)

The formal maximisation problem is: Maximise V(u) = V’(x, z, h(u), e(u))

(2.2)

subject to: Y – pxx – pzz – p(u)h(u) = 0

(2.3)

and T – e(u) – t(u) = 0

(2.4)

where V(u) is utility at location (u) and Y is household income. If this is maximised with respect to e(u) and location u, the following ﬁrst order conditions are obtained: ∂V’/∂e(u) – γ = 0

(2.5)

and –λh(u)[∂p(u)]/∂u – γt = 0

(2.6)

where the multipliers γ and λ are the marginal utilities, respectively, of leisure and income. The consumer’s optimal location is found by solving out for γ. h(u)∂p(u)/∂u = ∂V’/∂e(u)t/λ = pe(u)t

(2.7)

Urban Housing Models

25

where pe(u) is the monetised value of the marginal utility of leisure. At this optimal location, any move farther away from the CBD leads to a loss of leisure. The value of this loss is the marginal monetised value of leisure, pe(u), times the loss in leisure, –t. The optimal location conditions state that lost leisure is exactly offset by reduced housing costs such that utility is unchanged and welfare cannot be improved by moving. On the supply side, it is typically argued that the housing production function consists of two inputs, land and a composite housing structure capital input, and that the latter is perfectly elastic with respect to location. In elaboration beyond the simplest version of the model (where density is ﬁxed), density varies and this operates through the elasticity of substitution between these factors. At high residential densities, near the CBD, non-land is substituted for land but this intensity declines as production moves out into the suburbs. In a competitive land market, land rents are bid up nearer the CBD such that different land users will bid for land in different locations according to the trade-off that keeps them on the same iso-utility or iso-proﬁt function, creating a land rent gradient for the city as a whole. In equilibrium, price-taking households are at a tangency between the land rent gradient and their best (lowest) ‘bid-rent’ or marginal willingness-to-pay curve. Once income differentiation is included in the framework, and provided that higher income groups display an income elasticity of housing space demand that is greater than their price elasticity of commuting costs, then higher income groups will tend to locate towards the suburbs. Quigley (1979) summarises the outcomes of the basic model succinctly thus: ‘the principal conclusions of the monocentric model are: (1) residential densities decline with distance from the central place; (2) densities decline at a decreasing rate; (3) house prices decline with distance; (4) the land price gradient is steeper than the housing price gradient; (5) households with higher incomes locate further from the central place.’ (p. 39) Qualitative comparative statics associated with the model depend primarily on whether the city is assumed to be ‘open’ or ‘closed’ (Smith, Rosen & Fallis 1988). In the open set-up, there is an exogenous level of utility associated with living outwith the city. Should utility fall below a certain level, there will be outward migration and house prices will fall to restore equilibrium. In a closed city, population is given, while the level of utility is endogenous. McDonald (1997) runs through a series of comparative static exercises utilising a model based on Fujita’s contribution (1987). A decrease in the agricultural bid rent is demonstrated to decrease all bid rents, expand the city radius,

26

Housing Economics and Public Policy

increase average occupied household space and decrease population density within the old boundaries of the city at all distances. Increased population is shown to expand the CBD (to accommodate increased employment), increase equilibrium land rents, expand city size, increase population density and reduce average household space consumption.

Extensions, evidence and criticisms There have been many extensions proposed to the basic model. These have been aimed at relaxing certain assumptions in order to provide more richness or realism into the framework, or, alternatively, to try to take better account of empirical regularities at odds with the basic tenets of the model. Straszheim reviews many of these extensions, judging them valuable, though bought at the cost of ‘added complexity in modelling location and investment equilibria over time’ (Straszheim 1987, p. ·730). Muth (1985) also highlights this loss of tractability. Anas, Arnott & Small (1998) assess the access-space model in a broader historical context (see also Maclennan 1982) and ask what it can tell us about the long-term trends of urban change. They consider the conclusions drawn from the monocentric model against the changing patterns of centralisation and decentralisation, and clustering and dispersal of economic activities evident in modern cities, the phenomena of suburbanisation and the emergence of ‘edge cities’ (polycentric employment and activity nodes). The basic model cannot explain polycentrism (although Fujita’s 1987 framework incorporates both sub-centres and the labour market). However, it may shed some light on suburbanisation. Population density gradients have empirically declined over time in cities across a wide range of economic conditions. While this may be consistent with rising incomes and falling transport costs, Anas et al. (1998) note that commuting impacts more in lost leisure time for higher income groups. However, it has been difﬁcult to isolate and fully specify transport costs. Evidence suggests that within many cities the employment density gradient is larger than that for population but has been falling faster than the population gradient, and this can be construed to weakly support the hypothesis that jobs have been following people. Other possible explanations of decentralisation take us further away from the basic model. These include ‘ﬂight from blight’, racial discrimination, negative neighbourhood externalities created by poor areas, and Tiebout mechanisms (Anas et al. 1998, p.·1439).

Urban Housing Models

27

I have already covered a number of the typical criticisms of this class of models. These in sum relate to: the need for restrictive assumptions to deliver tractable results and comparative statics; consequent incapacity to deal with durability, dynamics and short-run market behaviour; and difﬁculty in dealing with certain patterns of evolution of real world cities, particularly with respect to decentralisation and polycentrism. In response, it may be argued that the inability of such models to deal with short-run dynamics reﬂects the fact that they seek to provide a general equilibrium account of longrun trends in urban land use and market outcomes – although they do not provide an altogether satisfactory account here either. A more fundamental criticism relates to the underlying and enduring relevance of this class of model. In other words, does it adequately address the important questions about urban housing markets (Maclennan 1982; Straszheim 1987); is it ﬁtfor-purpose? Urban housing market choices do not only rely on workplace accessibility and locational amenity but also reﬂect choices to do with neighbourhoods, housing product differentiation, the pervasive role of the state in the land market, urban transport structures and local ﬁscal bundles. The insights provided by search and bidding models, models that account for durability and product differentiation and by game theory approaches may be at least if not more important than highly reductionist spatial competition models in understanding housing processes. The slow adjustment, inherent lags and disequilibria tendencies of housing markets articulated by Whitehead & Odling-Smee (1975) present a very different (European?) conceptualisation of the housing market. Maclennan & Gibb (1994) suggest that the pervasiveness of policy impacts and the inﬂuence of history that play a central role in real urban housing systems cannot be adequately addressed within the access-space framework. One can also question whether the urban level is the appropriate scale for analysis as compared with lower levels such as the neighbourhood (Grigsby et al. 1987). A long-run urban-scale equilibrium construction with no role for the state or these other features simply misses too many of the key dimensions of housing markets. In the next section, I consider one of these dimensions: the dynamics of the existing housing stock at intra-urban levels through ﬁltering processes.

Filtering in urban housing markets Filtering models, with their focus on process and dynamics, and the relationship between new, rehabilitated and existing second-hand stock, are often far

28

Housing Economics and Public Policy

removed from the long-run equilibrium outcomes of the new urban economics (although see Olsen 1968, which I discuss below). This type of model has often been the basis for applied economic work and, of course, has a central and controversial place in housing policy debates. In this section, I brieﬂy review the development, differentiation and assessment of ﬁltering processes and models and point to their continuing relevance (the earlier literature is reviewed in Maclennan 1982, pp.·22–35). Homer Hoyt (1939) provided the ﬁrst study that might be labelled ‘ﬁltering’. Triggered by growing incomes and a preference for newness among high-income groups, Hoyt suggested that wealthier households would move out to newly developing suburbs. The formerly occupied neighbourhoods would in turn be replaced by the next socio-economic group, setting off a series of vacancies and moves down a price-income and sub-market hierarchy (Maclennan 1982). Apart from the strong implicit and explicit assumptions in this stylised model, addressed by Maclennan and responded to by a series of authors such as Grigsby (1963), there were also important policy ramiﬁcations. The ﬁltering process in the form presented suggested that promoting private market development of housing for high-income households could improve housing quality for all households throughout the housing system and implied that there was a case for subsidising such new development. Lowry (1960) argues that urban housing policy has been ‘haunted’ by the implications of this approach (p.·362). The crude or laissez-faire ﬁltering model has been comprehensively criticised (Robinson 1979; Maclennan 1982; Quigley 1998). First, it is often not well deﬁned (Lowry 1960). Second, it rests on assumptions that can be challenged empirically, for instance about the distribution of income and the location of income growth, as well as assumptions about the behaviour of non-moving households (Maclennan 1982). Third, as Lowry puts it, the ‘effectiveness of ﬁltering as a means of raising housing standards hinges on the speed of value-decline relative to quality-decline’ (1960 p.·364). If depreciation occurs more quickly than unit prices fall (as much evidence suggests) then the beneﬁcial ﬁltering process cannot occur. Fourth, as Quigley (1998) indicates, supply behaviour is endogenous. The ﬁltering of dwellings, therefore, will not just depend on incomes and preferences on the demand side and on physical depreciation on the supply side but also on the regulatory environment and cost structures faced by suppliers and the heterogeneous demands of real households (p. ·xiv). Fifth, from a policy point of view, Maclennan emphasises there may be ethical concerns about the subsidy targeting implied by the model. It may exacerbate decentralisation, and the

Urban Housing Models

29

ﬁltering may take a very long time to work through the system (especially given the impacts of trading-down within the system). Grigsby (1963) proposed a different interpretation of the ﬁltering concept. His concern was with ﬁltering as an outcome, and he sought to develop a better understanding of the empirical relationships between the stock of houses, households and moving decisions. Grigsby developed a matrix model where household moves are observed across housing quality or price ranges. As the market matrix evolves over time, households move around the system and this in turn has reinforcing effects, shifting the income mix and the average value of entire neighbourhoods. Using this approach, the study of wider dynamics becomes possible, for example, the knock-on effects, spillovers and backwashes that affect housing neighbourhoods after exogenous shocks take place or when, for instance, regeneration or rehabilitation takes place (Maclennan (1982) presents a nice summary of the Grigsby approach). From a very different perspective, Olsen (1968) addresses ﬁltering through his competitive model, which utilises the device of an unobservable unit of housing ‘service’. Olsen argues that ﬁltering represents changes in the quantity of housing stock, such that ﬁltering up (down) can only take place if the quantity of housing stock has increased (decreased). Filtering is a process by which the quantity of housing service yielded by a dwelling adjusts to the pattern of demand as a result of competitive proﬁt incentives (pp.·615–16). For Olsen, ﬁltering is an essential market mechanism that allows proﬁtmaximising housing producers to adjust supply and maintenance behaviour in order to move the competitive market toward equilibrium in terms of the price per unit of housing service. In fact, use of the ﬁltering approach has been hampered by a lack of clarity in how the concept should be deﬁned. Grigsby et al. (1987, Chapter 4) identify ten different approaches to ﬁltering in urban housing markets: ‘Depending on the deﬁnition, ﬁltering is either a process or set of outcomes and, if the latter, could involve, in differing combinations, change or no change in: occupancy, occupants’ incomes, value or rent of dwelling units, price per housing units or neighbourhood environs … It is not surprising that the literature is replete with controversy over whether ﬁltering occurs or works.’ (p.109) Galster & Rothenberg (1991) also distinguish between different forms of ﬁltering, which they consider within a quality-differentiated housing sub-market framework. First, when relative prices fall in a given sub-market, some

30

Housing Economics and Public Policy

households previously in lower quality sub-markets may now be willing and able to trade up to the now cheaper higher quality strata. This can be called price ﬁltering. Second, falling relative prices within a sub-market will also encourage some owners to change the quality of their dwelling to improve its rate of return, known as dwelling ﬁltering. Third, if the converted dwellings are spatially concentrated, households in such neighbourhoods may passively ﬁlter as a result of the spatial externality on the overall dwelling bundle value. Fourth, beyond the short run, these price-induced changes will lead to knock-on ﬁltering-like dynamics in related sub-markets through changes to prices, household moves and quality conversions. We can see elements of the ﬁltering approach in many urban housing models (Sweeney 1974; de Leeuw & Struyck 1975; Ingram 1979; Anas & Arnott 1991; Rothenberg et al. 1991; McDonald 1997). Discussion of these processes continues (Somerville & Holmes 2001). The inherent value of the concept of ﬁltering, both as process and outcome, makes it an important tool for urban housing market research. However the true potential of the concept remains to be fully realised.

Simulation models of urban housing One approach taken to capture more of the richness and interdependence of real housing markets has been the development of computer simulation models. They use parameter values for equations based on a mix of econometrics, good practice and judgement, to build more or less sophisticated urban housing models. These can in turn be used to model policy changes, and provide insights on housing market dynamics. These models can and have been widely criticised for their limitations (e.g. Maclennan 1986a; Gibb 1989) but recent developments, both in terms of the quality of the economic modelling they embody and computing power, have greatly improved their usefulness. There are still problems with this type of model, but their practical relevance and continued improvement suggest that they will become increasingly valuable.

Modelling developments Simulation models of the housing market provide a framework in which it is possible to consider the dynamic short-run path of market processes. The original 1970s and 1980s models, American in origin, were extensively used to forecast the effects of introducing housing allowance programmes and other tax reforms (early versions were reviewed by Ingram 1979). The two

Urban Housing Models

31

urban simulation models associated with this period were the UI (Urban Institute) model (De Leeuw & Struyck 1975) and the National Bureau (NBER-HUDS) model (Kain & Agpar 1985). Both models sought to allocate households to dwellings, while incorporating an explicit supply side. The UI model, for instance, consists of households, differentiated by age and race, which are allocated intertemporally to dwelling types across six residential zones (with differentiated workplace accessibility, and rents). Supply is set to maximise proﬁts and households maximise utility such that the different actors search for equilibrium outcomes in the markets for existing dwellings and new build. The UI model embodied certain ﬁltering processes. Attempts to allow choice and maximising behaviour within an (eventually) equilibrating framework can be seen through all of the successive models of this class. Smith, Rosen & Fallis (1988) argue that this type of model involves too much of a black box approach, where model parameters are not econometrically-deﬁned but: ‘simply chosen to ensure realistic model outcomes (and) are often ad hoc rather than based on economic behaviour.’ (p.·39) However, parameter estimates are likely to come from econometric work undertaken in other studies. Furthermore, it can be argued that these models have a different purpose and should be assessed in terms of their capacity to aid planning, forecasting, and policy development. Anas and Arnott, in a series of papers, have gone a long way to provide more secure theoretical foundations for urban housing simulation models. Anas & Arnott (1991) develop a robust dynamic discrete choice equilibrium model of the housing market that allocates different types of consumers to different housing types, clears the market and has an integrated supply side (including investors operating with perfect foresight). It is speciﬁcally designed to be useful for simulations and policy experiments. Anas & Arnott (1993c) further extend their work by developing and testing a housing market simulation model for Chicago. The key choices in the model concern the household’s choice of location, house type, quality and tenure status. The model demonstrates the feedback and second-round effects of speciﬁc policies on different parts of the housing quality hierarchy (and consequently on different income groups). In their 1994 paper, Anas & Arnott go on to test their model in four other US cities, examining the impact of housing allowances and construction subsidies. They ﬁnd that targeted demand-side allowances are more effective for the targeted groups

32

Housing Economics and Public Policy

than construction subsidies. This work is extended in 1997 by looking at the welfare effects of taxes and allowances on low-income households and the landlords providing their housing. Meen’s (1999a) paper has applied and extended urban-level discrete choice modelling of housing in a UK context, through the development of a model of the London and South East housing market. Meen starts from the argument that declining demand for public housing reﬂects, among other things, cumulative processes of decline as the least advantaged are left behind in the process of economic change and relocation. His work suggests that nonhousing social and neighbourhood factors need to be employed to capture, among other things, spatial externalities, and highlights three modelling implications. Population patterns vary within the region of study, implying different rates of new household formation and housing requirements. Second, the factors that drive household relocation include income changes, local social ties, the desire for a better neighbourhood, larger houses and changes in personal circumstances. Third, in-migration at the regional level must be speciﬁed and modelled.

Case study: the Glasgow citywide model of housing demand Recent work in Scotland has built on Meen’s study (and other work carried out on the South East of England (Meen & Andrew 1999)). A simulation model has been developed to forecast social housing demand in Greater Glasgow, where Glasgow is partitioned into three parts (North and South of the City and its suburbs). This model explicitly involves a tenure choice between social renting and owner-occupation (private renting is treated as part of the non-owning sector and subsumed within the analysis for forecasting purposes). The study (Gibb, Meen & Mackay, 2000) was designed to provide demand estimates for social housing under a range of assumptions about migration, neighbourhood and income change, taking full account of the economic choices of households, and the drivers of housing demand. Housing demand in each of the areas comes from three sources: migration into and out of the overall Glasgow area, new household formation, and moves among the three Glasgow areas of the existing population by tenure and by housing type. Each process is modelled. The existing residents housing choice model suggested that mobility decisions are based on demographics and income, and that tenure and locational choices are a function of neighbourhood quality, incomes, costs and the previous location of households. The supply side is modelled more mechanistically. For each sector, current supply is the sum of last period’s supply plus new investment by the

Urban Housing Models

33

construction sector, plus net transfers into the sector (which could be negative), minus demolitions and also minus long-term vacancies that reduce the effective stock. The model clears the market through supply adjustment (which essentially assumes that real house prices are constant). The model was estimated and then applied using a 1996 base, and projects forward to 2009 for the City of Glasgow and its suburbs (termed ‘outwith’ in the tables). Table 2.1 summarises the proportionate changes in demand, provides a combined total for the City of Glasgow and makes a correction for private rented housing, in order to calculate social rented demand. The estimates presented show the difference between the 1996 ﬁgures and the 2009 projections, in the absence of intervention.1 The numbers suggest that owning households in Glasgow will increase from 115 000 to just below 150 000. In the same period renting households (social and private) will fall from 155 000 to just over 130 000. In the suburbs, owning households will grow from more than 262 000 in 1996 to more than 346 000 in 2009. Suburban renters will decline from slightly more than 176 000 to a little more than 144 000. Table 2.1 suggests that social housing demand will fall in Glasgow by 18.1% of the 1996 level. At the same time, owner occupation in the City will increase by 30.2%. This reﬂects an overall increase in households in Glasgow of 3.7% from 1996 to 2009. For the area outwith Glasgow, social housing demand will fall by 19.1%. Owner-occupied demand in the outwith area will rise by 30.9%. Overall, households in the area outwith the City will increase by 11.2%.

Scenario simulations A scenario simulation alters the parameter value of one of the driver variables in the system so that one can trace the longer-term impact of the change by contrasting the outcome with the estimates reported above.

Table 2.1 Base housing demand change 1996–2009. Tenure and location

Absolute change (number of households)

Percentage change in households

Glasgow owning Glasgow social renting Outwith owning Outwith social renting

+34 681 –25 667 +81 104 –32 135

+30.2 –18.1 +30.9 –19.1

34

Housing Economics and Public Policy

Two scenarios are brieﬂy reported:

•

an improvement in neighbourhood quality in Glasgow relative to its suburbs

•

an increase in incomes within the wider Glasgow housing system.

Neighbourhood quality is proxied in the model through a geographicallybased deprivation index. In scenario one, the deprivation scores of Glasgow relative to its suburbs are changed to mimic a change in the relative quality of the City’s neighbourhoods. By examining the deviation from the base estimate benchmark it is possible to assess how tenure and location choices would shift as a consequence. Table 2.2 compares the original 2009 estimates with those of the model assuming a 5% improvement in the deprivation scores of the City. The simulated percentage change is a large one but the message is clear. Improving neighbourhood quality in this way should lead to a signiﬁcant increase in owner-occupation, primarily from owners moving into the City from the suburbs. However, some suburban renters would seek to move into owneroccupation within Glasgow as well. In scenario two, the assumed income change affects all of Greater Glasgow. This is more plausible than assuming that only the administrative jurisdiction of the City would beneﬁt from income changes. We increase average incomes after 2002 by 10% and compare the new housing demand conﬁguration that occurs with the original estimate for 2009. Table 2.3 indicates that increased incomes would increase demand for owning in Glasgow and in the suburbs. Renting demand falls substantially in both Glasgow and its suburbs. Income effects would appear to play their usual role of increasing owner-occupation but, in addition to reducing rental demand, Glasgow households are using their higher incomes to purchase wider aﬁeld, reﬂecting supply constraints in Glasgow. Table 2.2 Simulation impact of 5% improvement in Glasgow’s deprivation scores in terms of overall households by tenure. Location and tenure

2009 [central] (1)

5% improvement in score (2)

Deviation (2–1)

Glasgow social renters Glasgow owners Outwith social renters Outwith owners

116 508 149 607 136 380 346 334

108 805 175 325 132 599 333 390

–7 703 25 718 –3 781 –12 944

Urban Housing Models

35

Table 2.3 Simulation impact of 10% increase in incomes after 2002. Location and tenure

2009 [central] (1)

10% increase in incomes (2)

Deviation (2–1)

Glasgow social renters Glasgow owners Outwith social renters Outwith owners

116 508 149 607 136 380 346 334

108 480 153 320 126 355 362 790

–8 028 3 713 –10 025 16 456

The Citywide model is the ﬁrst serious attempt in the UK to build a simulation forecasting model of the entire housing system for a metropolitan area, explicitly including the social rented sector. In many ways, it is much less developed than the Anas & Arnott model for Chicago but it does incorporate several desirable features of such models – in particular a defensible multilevel discrete choice model of housing choice, and the use of well-determined models of regional migration. There are also important weaknesses with the model as it stands. The supply side is under-developed. There is only an implicit residual role for the private rented sector, and there is no explicit modelling of transition probabilities for household change (such as the transition from a pensioner couple to a single pensioner household). The crude spatial nature of the model must also be acknowledged. Notwithstanding these limitations, the results show the potential of this class of model to support city level planning and policy development.

Conclusions In conclusion, I return to the issue of the comparative strengths of the different approaches to urban housing market modelling before considering the wider theme of future developments in urban housing market research. The trade-off model of new urban economics raises three fundamental problems as an approach seeking to explain housing market behaviour. First of all this type of model is highly reductionist, and relies on the concept of unobservable housing services and other simpliﬁcations to make it tractable. In doing this, such models remove many of the important elements or aspects of the urban housing system of interest (e.g. the role of the housing stock, its history, the role of the state, the possibility of intra-urban housing submarkets, or any focus on adjustment and disequilibrium). Second, even on its own terms, this type of model has difﬁculty coping with certain empirical

36

Housing Economics and Public Policy

regularities one might expect it to explain, such as suburbanisation and the development of edge cities. Third, by excluding all other dimensions except the spatial, this class of model removes from the ﬁeld alternative conceptual starting points such as search and bidding, bargaining, imperfect competition, housing market segmentation and product differentiation. However, there are important and enduring contributions that access-space models have made. First of all they make explicit the link between housing and transport within urban systems. If one rejects the mechanics of the access-space model that does not mean that one can afford to neglect transport and other land use issues. Second, they are clearly correct in emphasising employment and general accessibility factors as important drivers of housing demand. Third, these models do provide a well-deﬁned benchmark long-run spatial equilibrium position, from which deviations and different trajectories can be considered and explained. Filtering models as a class have the potential to provide considerable insights once we accept the importance of housing market segmentation, sub-markets and neighbourhood structures. However, for several decades their use has been hamstrung by deﬁnitional problems and by earlier policy baggage. Recent work by Galster & Rothenberg (1991) has been valuable in addressing these problems. Filtering models provide a very different set of modelling tools for housing market analysts. The ideas of household and dwelling succession and change, depreciation, upgrading and market dynamics are immensely useful for applied economic analysis of housing and neighbourhoods. Simulation models embody elements of both trade-off and ﬁltering models, and are built for very practical purposes. They have been shown to be potentially useful tools for policy development and planning purposes. However, these models are only as good as the economic and econometric foundations they are built on and neither the access-space nor the ﬁltering frameworks provide the basis for a comprehensive urban housing model. This should not be surprising. Filtering and access-space models are designed to address both different and speciﬁc types of questions and represent different methodological forms of explanation. Maclennan & Gibb (1994) proposed a forward direction for UK housing economics. If one wishes to understand urban housing markets, then modelling disequilibrium processes and accounting for the role of key institutions will be at least as important as seeking to deﬁne equilibrium outcomes. In understanding markets for urban housing, it is important to have a clear conceptual and empirical understanding of the system under investigation:

Urban Housing Models

37

‘It is critical to understand its evolution. A modelling of the heuristics of behaviour in the context of limited information is required. And it is critical to understand the institutions which emerge, adapt and dissolve as the system unfolds’ (1994 p. ·4, original emphasis). The challenge for the future development of spatial models of the housing market is to incorporate these features in a rigorous but relevant model that can be sensibly operationalised and disseminated. The recent work by Geoff Meen and colleagues is a possible way ahead (Meen 1999a, 2001; Meen et al. 2001; and see the chapter by Meen in this volume). Here, the focus is on the explanation of movement of households, tenure change and the growth and decline of urban areas. Important themes include the processes of cumulative causation, exclusion and the factors that shape intra-urban housing choices by existing households. This work is at an early stage but it is suggestive of a promising way forward for UK and wider housing market research.

Acknowledgements These are due to Steve Malpezzi, Tony O’Sullivan and George Galster for comments on an earlier draft. The chapter also beneﬁted from the comments made by participants at the University of Glasgow workshop for the book, held in November 2001.

Note (1) In other words, in the absence of any response by planners in terms of land release as a result of observed demand levels or change thereof.

3 Segmentation, Adjustment and Disequilibrium Yong Tu

Introduction Housing is a complex commodity. Of special interest here is how this complexity creates housing market segmentation and the market consequences of segmentation, which I discuss in the next section. I move on from this to discuss the equilibrium and disequilibrium properties of a segmented housing market, and then the issues involved in empirical identiﬁcation of housing sub-markets. I conclude with a discussion of the wider implications for the study of housing economics of rejecting equilibrium as a starting point in modelling the housing market.

The causes and consequences of housing market segmentation Housing as a complex commodity can be considered in terms of three dimensions (see the discussion in Galster 1996): spatial immobility, durability and heterogeneity. Spatial immobility implies that characteristics associated with a location are inherent in the bundle of attributes of a dwelling found at that location. These locational attributes include the socio-economic status of the neighbourhood, its physical conditions, as well as wider notions of accessibility to any desired destinations, in terms of jobs, relatives and friends, private goods or public facilities. Combined, all of these attributes contribute to differences in housing quality and housing prices across locations. Housing is also durable, and typically has a relatively high purchase price involving the need for extensive mortgage borrowing. It is also common for dwellings to experience long run price appreciation. The investment aspect

Segmentation, Adjustment and Disequilibrium

39

of owning a home, therefore, is likely to make housing a principal personal sector ﬁnancial asset. This also means that the housing market shares certain characteristics with the capital asset market (Maclennan 1982). On the supply side, a dwelling is modiﬁable. The quality of a housing unit can be either improved through modiﬁcation or reduced in value through depreciation over time at any location. Modiﬁed dwelling units form an important source of housing supply in any housing market. Heterogeneity of dwellings’ physical characteristics refers to the intrinsic variation found across housing types, sizes, ages, building materials, exterior or interior structures, and architecture designs, as well as different forms of land leasing. Heterogeneity generates variation in housing quality and house prices within a location, providing a homebuyer with a variety of dwelling choice alternatives. The multidimensional heterogeneity of an urban housing market gives rise to its likely segmentation. An urban housing stock that is immobile, durable yet capable of modiﬁcation, and heterogeneous, can be thought of as being segmented into a series of interrelated but distinctive housing product groups (Maclennan et al. 1987). The dwellings within each product group are more substitutable than the dwellings between the product groups and the degree of substitutability varies from group to group. At the same time, homebuyers are also heterogeneous. Past studies of housing choice behaviour (Quigley 1982b; Tu & Goldﬁnch 1996) show that would-be homebuyer preferences are determined and differentiated by, amongst other things, income levels, employment status and occupation, household composition, and previous housing experiences. A would-be homebuyer is likely to be interested in dwellings contained within only a few closely substitutable housing product groups. These product groups are likely to be spatially separated, implying the prospective buyer may only have imperfect knowledge regarding the pattern of spatially-dispersed dwelling units offered by sellers. Because housing is also one of the most expensive consumption and investment goods purchased by a household, it may be expected, for any would-be homebuyer, that the consideration given to (and evaluation involved in) purchasing a dwelling will be relatively intense. Search activity is required to inform the process of housing choice decision-making (Clark 1982a), which incurs both time and money costs. Maclennan (1982) points out that the housing purchase process involves a would-be homebuyer typically having not only to search and evaluate dispersed housing offers but also to search in order to secure home loan ﬁnance. It also requires bidding if there is more than one potential buyer for a dwelling unit (abstracting from the

40

Housing Economics and Public Policy

institutional details of house purchase in individual countries). Therefore, housing search may have three outcomes. First, a buyer ﬁnds the desired dwelling from a particular product group and is happy with the offer price. The transaction happens if there are no other competitors on the demand side.1 Second, the dwelling units offered by the product group are not acceptable to the potential buyer. A housing mismatch arises. Due to the slow adjustment of housing supply, to ﬁnd the desired dwelling, the potential buyer has to either continue searching among other groups or adjust their housing preference.2 Such mismatch creates high transaction costs for both buyers and sellers. For the buyers, it will create higher search costs and search time. For the sellers, they have to pay all the costs associated with retaining the dwelling unit. Market trade friction may therefore arise and force the seller to lower the offer price as a trade-off against an anticipated lower transaction cost. Third, where multiple potential buyers are interested in one dwelling unit, this gives rise to another form of housing mismatch. The bidding competition will eventually drive up the transaction price. In the absence of instantaneous housing supply adjustment, would-be homebuyers’ chances of obtaining desired housing attributes depend on ability and willingness to pay for them, as well as the ability to successfully negotiate the bid in the market (Wood & Maclennan 1982). The buyer who can afford the highest price will get the unit. This buyer may particularly be interested in certain attributes associated with the dwelling – for example, a dwelling close to good schools – and consequently be willing to pay a higher premium to obtain this housing attribute, increasing the locational attribute price among certain product groups. Differences across housing product groups in the price of a particular housing attribute (also deﬁned as hedonic price or implicit price; see DiPasquale & Wheaton 1996; Dale-Johnston 1982) give rise to topographicallybased housing sub-markets. A topographically-based housing sub-market consists of a cluster of housing product groups that may or may not have a contiguous geographical boundary. The price of an individual housing attribute has no statistically signiﬁcant differences within these product groups, but does have signiﬁcant differences across housing product groups belonging to different topographically-based sub-markets. This deﬁnition is supported by the empirical ﬁndings of Schnare & Struyk (1976), Goodman (1981), Maclennan et al. (1987), Goodman & Thibodeau (1998) and Bourassa et al. (1999). The theoretical discussions given by Tu (1997) and Maclennan & Tu (1996) also suggest

Segmentation, Adjustment and Disequilibrium

41

that the spatial patterns of topographically-based housing sub-markets may change over time. A different approach to sub-market deﬁnition is that of quality-based housing sub-markets, as deﬁned by Rothenberg et al. (1991). A quality-based housing sub-market consists of a sub-set of dwellings in an urban housing market. The dwellings in the group, though different, are viewed by demanders and suppliers as producing a similar level of quality. Thus, the urban housing stock as a whole can be considered as arrayed along a single quality dimension, across which exists a spectrum of sub-markets ranged from the highest to the lowest quality levels. Under this deﬁnition, dwellings within a sub-market are more highly substitutable than the dwellings between the sub-markets.3 A hybrid housing sub-market approach was proposed by Tu (1997). Here, the urban housing sub-market structure is conceived as a nested framework. Within the structure, the market is divided into topographically-based sub-markets, and inside each topographical sub-market, the dwellings are grouped into different sectoral housing sub-markets deﬁned by dwelling type. A similar approach was taken in Straszheim (1974) and, more recently, by Goodman & Thibodeau (1998). Maclennan (1982) points out that there are good reasons to believe that, due to the existence of housing sub-markets, disequilibrium characterises the urban housing market and may be pervasive. First, disequilibrium appears in the form of mismatch between demand and supply, and continuously purposive housing search creates housing sub-market trade frictions, ensuring that the housing market cannot clear its excess demands or supplies instantly. Second, the impacts of mismatch on market disequilibrium are reinforced by the slow adjustment of housing supply. Changes in demand for the dwellings in a sub-market occur more quickly than changes in supply. Therefore disequilibrium is likely to be persistent. These intrinsic disequilibrium features associated with an urban housing market suggest that any theory that hopes to provide a robust explanation of urban housing market dynamics or that seeks to provide a framework for evaluating housing policies must take the features of housing market disequilibrium into account. The conventional concept of neoclassical market equilibrium is particularly inappropriate in the urban housing market because factors such as transactions costs (including search costs) and

42

Housing Economics and Public Policy

market trade frictions are of more than usual importance. Economic models of urban housing will yield satisfactory prediction of real world phenomena only if they take such factors into consideration (Whitehead & Odling-Smee 1975).

A model of housing market disequilibrium This section draws on the preceding review and provides a theoretical framework from which housing sub-market equilibrium and disequilibrium properties are derived. The framework is a hybrid model of housing sub-market structure and represents a modiﬁcation and an extension of Tu (1997). Consider an urban owner-occupied housing market, and assume that this market is divided into two topographically-based housing sub-markets (TiSubmkt, i=1,2). Within each topographically-based housing sub-market, there are two quality-based housing sub-markets (Qij-Submkt, i = 1,2 and j = 1,2). In each quality-based housing sub-market, the dwellings not only share the same spatial characteristics (for example, neighbourhood and public amenities) but are also highly substitutable. The housing economics literature provides evidence with respect to a would-be homebuyer’s housing choice behaviour. First, housing search is a requirement and search costs limit a would-be homebuyer to seek housing in selected topographical areas (Wood & Maclennan 1982). Second, a would-be homebuyer’s housing preference, for example, the desire to live close to friends or workplace, may conﬁne him or her to certain topographical areas (Tu & Goldﬁnch 1996). Third, ﬁnancial affordability constraints may also restrict opportunities to choose a dwelling located in certain expensive areas (Tu & Goldﬁnch 1996). Therefore it appears that a would-be homebuyer often examines housing in a few limited areas, and housing choice decision-making processes start from a favoured area. Within this area, as a rational buyer, the potential purchaser will continue to search across dwelling units, with particular attention paid to those dwellings satisfying housing quality requirements. For this would-be homebuyer, the dwellings of interest form one quality-based housing sub-market because the purchaser evaluates those dwellings that are highly substitutable (providing approximately the same level of housing

Segmentation, Adjustment and Disequilibrium

43

quality). The searching process stops if and when a dwelling unit from the area can be found that can generate the maximum household utility. Otherwise, the supply constraints will drive the potential purchaser to switch to another favoured area to keep searching (which I shall call re-choice). Supply constraints mean that housing supply does not respond to demand changes instantaneously, due to the slow adjustment of housing stock in the area and the lack of vacant land in built-up areas. The hybrid model of housing sub-market structure presented in this section reﬂects this housing choice decision-making process. In an urban area, a well-deﬁned hybrid housing sub-market structure may assist the prospective homebuyer to gather more housing information and hence to reduce market trade frictions. It may also aid housing suppliers to identify development opportunities from the demand–supply gaps within each sub-market. The operational process of this urban housing market is a dynamic matching process between the would-be homebuyers and the sellers (see Fig. 3.1). In the ﬁgure, ‘T-choice’ and ‘Q-choice’ refer respectively to the choice between topographically-based housing sub-markets and between quality-based submarkets. The current state of an urban housing market is a group of k points in a threedimensional space ΓK,

Re-choice T-Submkt

Re-choice Q-Submkt

Q11-Submkt T1-Submkt

Extra housing demand in-flow Dt

T-choice

QChoice

Re-choice T-Submkt Exit

Re-choice Q-Submkt

Re-choice T-Submkt

Re-choice Q-Submkt

T2-Submkt

Re-choice T-Submkt

Keep searching Trading

Q12-Submkt

Trading Keep searching

Q21-Submkt

Keep searching Trading

Q22-Submkt

Trading Keep searching

QChoice

Re-choice Q-Submkt

Fig. 3.1 Urban housing sub-market operational system.

44

Housing Economics and Public Policy

where k is the number of housing sub-markets. X and Y are the non-negative real axes. Z is the positive real axis. For example, in Fig. 3.1, there are 6 housing sub-markets (k = 6). The current state of the housing market at any time point can be represented as a matrix W(t) illustrating the 6 points in the space ΓK.

In the matrix, wij, i = 1,2 and j = 0,1,2, indicates the current situation of either a respective topographically-based sub-market (for j = 0) or a quality-based sub-market (for j = 1,2). Each situation is determined by three sub-market variables: the number of housing units demanded (d), the number of housing units supplied (s) and the price (p). Each variable is, in turn, deﬁned on the axes of X, Y and Z. As long as the urban housing market is not in equilibrium, this group of 6 points (W(t)) moves in the deﬁned space. Mathematically, the dynamic process can be expressed as:

(3.1)

(3.2)

(3.3)

where i = 1,2, j = 0,1,2. a, c and b are the extra housing demand, supply inﬂows to the respective sub-market, and the number of dwellings transacted at time t. Obviously, at the level of the topographically-based sub-market, ai0, ci0, bi0, di0 and si0 are the sum of the respective variables at the level of quality-based sub-markets nested in the relevant topographically-based sub-market. χ is a coefﬁcient reﬂecting changes to the sub-market house prices in response to the difference between sub-market demand and supply. Changes to sub-market housing demand, supply and price, at any time t, are determined by the urban housing market situation (W(t)) as indicated in equations (3.1)–(3.3).

Segmentation, Adjustment and Disequilibrium

45

The necessary and sufﬁcient conditions for an urban housing market achieving equilibrium are: (3.4) where i = 1, 2 and j = 0, 1, 2

(3.5)

The above equilibrium conditions are deﬁned as the stock-ﬂow market equilibrium (Weibull 1983, 1984). Three important implications can be drawn from the stock-ﬂow equilibrium conditions. First, equations (3.3) and (3.4) imply that, when the housing market achieves stock-ﬂow equilibrium (equilibrium at both topographically-based sub-markets and quality-based sub-markets), there is no price movement in each sub-market. Second, even in the equilibrium, there exists unsatisﬁed demand (homebuyers keep searching) as well as unsold dwelling units (some sellers search for a suitable buyer) in any of the housing sub-markets (equation 3.5). This phenomenon is a result of housing sub-market trade friction (TF) measured by equation (3.6). Third, the stock-ﬂow equilibrium in each of the quality-based housing submarkets is a necessary, but not sufﬁcient condition for the stock-ﬂow equilibrium in each of the topographically-based housing sub-markets. where i = 1, 2 and j = 0, 1, 2

(3.6)

A bigger value of TF implies a higher level of trade friction. The reciprocal of TF is interpreted as the housing sub-market transaction velocity (β). The seller’s average waiting time (q) in each sub-market can therefore be deﬁned by equation (3.7). (3.7) To derive the short-run and long-run housing sub-market equilibrium/ disequilibrium properties, four assumptions are needed.

46

Housing Economics and Public Policy

Assumption 1 An urban housing market is assumed to be a free market. There is no regulated price control and changes in sub-market housing demand and supply are the only cause of house price ﬂuctuations.

Assumption 2 The market is assumed to be in a stable socio-economic environment. Therefore, the extra housing demand inﬂow in to the housing market at any time t (Dt) is time-invariant.

Assumption 3 New additions to or deletions from the housing market are not considered. The only source of supply is via turnover in the second-hand housing market, and the turnover rate (ε) is a constant indicating a fraction of the occupied dwellings, that is, the difference between the sub-market dwelling stock (δ) and the number of dwellings on sale in the sub-market. To simplify the discussion, it is assumed that, at any time, there is only one consumer group in the market sharing the same housing preference distribution. Therefore, the subscript denoting an individual buyer is dropped from equations (3.8)– (3.11). These equations indicate a measurement of the choice probability of an individual buyer in choosing a housing unit from a sub-market. Given the hybrid housing sub-market structure (see Fig. 3.1), a nested multinomial logit model is used here to measure the probabilities (McFadden 1978).

where i, j = 1, 2

where i = 1, 2

(3.8)

(3.9)

(3.10)

where i = 1, 2

(3.11)

where (1–σ) is the coefﬁcient of the inclusive value representing a buyer’s

Segmentation, Adjustment and Disequilibrium

47

preference for the nested quality-based sub-market as a whole, X and Y are a vector of variables that may inﬂuence a would-be homebuyer’s housing choice (X indicates physical dwelling conditions, Y indicates dwelling related spatial attributes), α1 and α2 are negative. Pr00 indicates the probability of quitting the urban housing market after searching. Pri0 indicates the probability of entering a topographically-based sub-market while Prj/i indicates the probability of entering a quality-based sub-market after entering the respective topographical sub-market. After a would-be homebuyer (a housing searcher) enters a quality-based sub-market at any time t, he or she either ﬁnds a dwelling and a housing transaction occurs, or he/she has to leave the sub-market, to adjust his/her preference and to make a re-choice between the topographical sub-markets or the quality-based sub-markets (Fig. 3.1).

Assumption 4 In each quality-based sub-market, the re-choice rate ( ) is a constant indicating the fraction of housing searchers in the jth quality-based sub-market nested in the ith topographically-based sub-market who join the re-choice ﬂow after searching in the respective quality-based sub-market, where k = 2 indicates that the would-be homebuyer makes a re-choice between the topographical sub-markets and k = 1 indicates that the would-be homebuyer makes a re-choice between the quality-based sub-markets. Based on the above four assumptions and equation (3.7) that deﬁnes the housing sub-market trade friction, the dynamics of urban housing sub-markets (a, b, and c) can be speciﬁed (as equations 3.12–3.17). In each qualitybased sub-market: aij(t, W(t)) = (3.12)

(3.13)

(3.14) where i, j = 1, 2. In each topographically-based sub-market:

48

Housing Economics and Public Policy

(3.15)

(3.16)

(3.17) Substituting equations (3.12)–(3.17) into the stock-ﬂow equilibrium conditions (see equations 3.4 and 3.5), a unique stock-ﬂow market equilibrium position is obtained (equations 3.19–3.23) for this hybrid urban housing submarket system (Fig. 3.1), if equation (3.18) is satisﬁed. The superscript ‘*’ used in the following equation denotes equilibrium and ‘t’ is dropped from the equations. The mathematical derivation of equilibrium is given in Tu (1997). where i, j = 1, 2.

(3.18)

Equation (3.18) means that the pre-condition for a hybrid housing sub-market system achieving its stock-ﬂow equilibrium requires that the total extra housing demand is bigger than the total extra housing supply in the housing market. The housing market is, therefore, demand-driven. The stock-ﬂow equilibrium is then achieved if the following equations are met (equations 3.19–3.23). (3.19)

(3.20)

(3.21) and (3.22)

(3.23)

Segmentation, Adjustment and Disequilibrium

49

where i, j = 1, 2. A set of static and dynamic equilibrium and disequilibrium properties in an urban housing market can be derived from the above equilibrium equations.

Static equilibrium properties First, the sub-market equilibrium housing price is determined by four factors (equation 3.21). (1) The price is positively related to the bundle of attributes associated with a dwelling. (2) The price is positively related to a homebuyer’s preference for each housing attribute in a sub-market. In equations (3.8) and (3.9), α3 and λ indicate a would-be homebuyer’s preference for a housing attribute. (3) The price is negatively related to the seller’s average selling time in a sub-market. Therefore the sub-market housing price is negatively related to the level of housing sub-market trade friction (see equations 3.6 and 3.7). (4) On the right-hand side of equations (3.22) and (3.23), the numerator is the sum of extra sub-market housing supply and the total number of housing searchers who join the re-choice ﬂow in the respective sub-market. A large sum at equilibrium implies that the sub-market has a higher trade friction. Therefore the right-hand side of equations (3.22) and (3.23) indicates the relative trade friction of a sub-market to the average. Bringing equations (3.22) and (3.23) together with equation (3.21), it can be concluded that a sub-market with a relatively high level of market trade friction in an urban housing market will have a lower level sub-market housing price at equilibrium, given other factors remaining the same. Second, increasing the housing demand inﬂow to an urban housing market will generally increase house prices in each sub-market (bringing equation 3.23 into equation 3.21). Any socio-economic factor that may change the demand inﬂow will have signiﬁcant impacts on sub-market house prices. Third, equilibrium sub-market housing demand and supply are positively related to the sub-market housing stock and the sub-market turnover rate,

50

Housing Economics and Public Policy

but negatively related to the sub-market transaction velocity. A faster transaction velocity deﬁned by βij in equation (3.13) can reduce the numbers of those searching in the sub-market, giving rise to a lower level of equilibrium sub-market demand and supply. Higher βij means that, at any time, a larger proportion of buyers can ﬁnd a matched dwelling unit if supply exceeds demand, or a larger proportion of sellers can ﬁnd a matched buyer, provided that supply is less than demand. In other words, the market trade friction is lower (equation 3.20), which leads to a lower level equilibrium house price (equation 3.21).

Dynamic disequilibrium properties Dynamic sub-market disequilibrium properties can be derived from a computer simulation based on equations (3.18)–(3.23). The details of the simulation are given in Tu (1997). These dynamic properties illustrate the feature of housing sub-market dynamic stock-ﬂow equilibrium stability. Stability analysis is an important part of equilibrium theory: many economic theories depend on the comparative statics of equilibrium. However, equilibrium properties make sense only if the underlying equilibrium system is stable. Four types of stability can be deﬁned in terms of degree. Global stability means that the market price can converge to equilibrium from any disequilibrium position. Neighbourhood stability means that the market price can only converge to equilibrium from a disequilibrium position located in the neighbourhood of an equilibrium position. Bounded instability is a mixture of stability and instability. Herein, market price follows an oscillatory movement around its equilibrium after it leaves an equilibrium position but still lies within the neighbourhood of the equilibrium. Instability means that the market will never return to equilibrium after once leaving its equilibrium position. Computer simulations reveal that an urban housing market stock-ﬂow equilibrium can take the form of a neighbourhood equilibrium, bounded instability or instability depending on the levels of substitutability between sub-markets. A highly segmented housing sub-market structure (or a low level of housing substitutability between sub-markets), which is determined by signiﬁcant differences in housing quality and housing attributes between sub-markets, will have a higher level of stability. The reason for this is that a highly segmented housing sub-market structure results in signiﬁcantly different choice probabilities for a would-be homebuyer within the sub-markets (equations 3.8–3.11). If a would-be homebuyer could not ﬁnd a satisfactory dwelling in a sub-market, he/she would have less chance to re-enter

Segmentation, Adjustment and Disequilibrium

51

other sub-markets and maintain housing search due to the high degree of segmentation. The average searching time for each would-be buyer is therefore shortened, which reduces the sub-market trade friction. The market, therefore, takes a shorter time to get back to equilibrium. The study of urban housing sub-market structure and its equilibrium properties is an important dimension of the analysis of urban housing markets and also provides a solid framework to understand the links between an urban housing market and the other sectors of an urban economy. However, submarket models can only be used in the analysis of housing and the urban economy if the market is carefully divided into sub-markets that are likely to show a discrete attribute price structure (Palm 1978). The next section introduces the key problem of sub-market identiﬁcation.

Urban housing sub-market identiﬁcation Although dividing the urban housing market into sub-markets raises numerous theoretical and methodological questions (Palm 1978), Goodman & Thibodeau (1998) point out that urban housing sub-market identiﬁcation has typically been performed on an ad hoc basis. A signiﬁcant difﬁculty with the ad hoc procedure for identifying sub-markets is that one cannot be conﬁdent that the resulting sub-markets are identiﬁed in an optimal way (Bourassa et al. 1999). The result is, to some extent, determined by the grouping methodology adopted. However, there are three criteria to apply in identifying urban housing sub-markets (adapted from Cliff et al. 1975): (1) Simplicity – a solution with few sub-markets is superior to a solution with many sub-markets. (2) Similarity – the housing attributes associated with a dwelling should be as similar as possible within a sub-market. In other words, the dwellings in a sub-market should have a high degree of homogeneity (or substitutability), while the dwellings in different sub-markets should be subjected to a higher degree of heterogeneity. (3) Compactness – dwellings located in contiguous areas should be more likely to be grouped into one sub-market than dwellings in more distant areas. A variety of housing sub-market grouping methodologies have been developed to identify either topographically-based housing sub-markets or quality-based housing sub-markets (Rothenberg et al. 1991; Watkins 1998).

52

Housing Economics and Public Policy

Almost all these methodologies, albeit different, are derived from the hedonic housing price model (this model is discussed in detail in the chapter by Stephen Malpezzi). The major limitation of using the hedonic price model is the lack of a solid theoretical basis for the identiﬁcation and the measurement of the associated housing attributes and the choice of functional form. The results obtained using the hedonic approach therefore may not be optimal or unique. In practice, the identiﬁcation of neighbourhood amenities has usually been based on considerations of data availability, and measurement based on the convenience of dealing with the data at hand, particularly for qualitative variables. Halvorsen & Pollakowski (1981) argue that economic theory does not suggest an appropriate functional form for hedonic price functions. But it is possible to develop a general functional form that incorporates all types of functional forms of possible interest as special cases. They suggest using the Box-Cox ﬂexible functional form for hedonic analysis and measuring the best performance with a goodness-of-ﬁt test. The major criticism of Halvorsen & Pollakowski’s method is that the advantage of adopting Box-Cox functional form is purchased at the expense of other important goals in hedonic analysis. Cassel & Mendelsohn (1985) point out that using the Box-Cox functional form may be problematic in several respects. First, the large number of coefﬁcients estimated with the BoxCox functional form reduces the accuracy of any single coefﬁcient, resulting in a poorer measurement of hedonic price. Second, the Box-Cox functional form is not suited to any dataset containing negative numbers. Third, the Box-Cox functional form may be inappropriate for predicting house prices. Fourth, the non-linear transformation used in the Box-Cox functional form results in complex estimates of elasticities, that are often too cumbersome to calculate. Despite criticism, the hedonic price model has proved the most popular method employed in housing sub-market identiﬁcation. Identifying topographically-based housing sub-markets generally takes two steps. First, the housing market is divided into market segments on the assumption that the dwellings within a segment are more homogeneous than the dwellings between segments. Typically, researchers have formed market segments based on prior expectations related to municipal boundaries, census tracts, school districts or racial divisions. In many cases, there has been no basis on which to assume that the dwellings are homogeneous in the areas

Segmentation, Adjustment and Disequilibrium

53

employed, and thus the resulting sub-market identiﬁcation may have been neither optimal nor unique. To solve this problem, Bourassa et al. (1999) propose a more systematic method to form market segments. They ﬁrst apply principal component analysis to a large dataset to extract a set of principal factors from a group of housing-related attributes. Then the factor scores are calculated for each dwelling and are used in cluster analysis to form homogeneous dwelling groups (that is, market segments). Although this methodology provides a comprehensive procedure for constructing housing market segments, the disadvantage is that dwellings that are located far apart may be grouped into one market segment, resulting in the ﬁnal sub-market identiﬁcation possessing a lack of compactness. This difﬁculty presents a research challenge to develop a more powerful grouping method that more effectively tackles the spatial aspects of housing market segmentation. Second, once market segments are identiﬁed, a hedonic price function is estimated for each segment to analyse hedonic house prices disparities across them. A housing segment becomes a housing sub-market if its hedonic prices are statistically signiﬁcantly different from the ones in the rest of the housing segments (Goodman & Thibodeau 1998). Statistical tests – for example, a Chow test, a weighted standard error test or an F-test – can be used to test if there are any signiﬁcant differences in the hedonic house prices between segments (see Schnare & Struyk 1976, and the review by Watkins 1998). Segments demonstrating no signiﬁcant differences in the hedonic prices are merged into a new segment, for which the hedonic function is then re-estimated. The procedure used in identifying quality-based housing sub-markets is usually based on the method proposed by Rothenberg et al. (1991). Again, dwellings within a sub-market are more substitutable than dwellings between sub-markets. Substitutability implies a functional ‘nearness’ for consumers and builders. The gradations of substitutability are given by utility differences perceived by a household, which provides the theoretical basis for deﬁning a single-dimensional index of the quality of housing services provided by a dwelling unit. Empirically, it takes two steps to divide an urban housing market into quality-based housing sub-markets. First, a hedonic function is applied to the sample to obtain the hedonic price for each dwelling-related attribute. A housing quality index for a speciﬁc housing unit can then be operationally deﬁned as the value of that unit calculated through the estimated hedonic function. Second, the dwellings are arrayed according to their quality index, and then the array is subdivided to form quality-based housing sub-markets. The

54

Housing Economics and Public Policy

disadvantage of this method is that households differ markedly in their taste for housing. Different households are likely to disagree about the pattern of substitutability. Therefore, no quality index would be unanimously valid for any two households. Housing sub-market identiﬁcation has important ramiﬁcations for housing market analysis. First, Watkins (1998) argues that the clear evidence of housing sub-market existence from a large range of studies implies that failure of housing economists to adopt the housing sub-market as a central concept in housing market analysis may lead to spurious conclusions. Second, a welldeﬁned urban housing sub-market structure is particularly important for applying the hedonic technique to property valuation (Adair, Berry & McGreal 1996; Watkins 1998), in the construction of property indices (Berry, Chung & Waddell 1995), for housing market planning (Jones & Watkins 1999) and in the evaluation of urban policy initiatives. However, there is, in fact, nothing intrinsic to the notion of a housing submarket that leads one to conclude that the use of any of the methodologies discussed is superior to the others. Each methodology requires assumptions about, for example, how market segments or housing quality are deﬁned, perceived and agreed upon by participants in a housing market. The dynamic features of housing sub-market structure also suggest that it may be necessary to re-classify or re-identify housing sub-markets after a certain time period, for example, every 3–5 years.

Conclusions Heterogeneity, locational ﬁxity and durability underpin the distinctive nature of the housing commodity, and, together with heterogeneous housing demand, imply that the urban housing market should be considered as segmented. Housing sub-markets are likely to exist, but there will be linkages between them. Any changes in the demand, supply and price within a sub-market may lead to changes to demand, supply and price in the remaining sub-markets, but the impacts will be stronger in those that are closer substitutes to the sub-market initiating the change. Therefore, changes beginning in one sub-market will generate repercussions in others, and in time these may re-establish the original interrelationship between the submarkets (Rothenberg et al. 1991). Disequilibrium characterises an urban housing market. The urban housing sub-market concept lays a foundation upon which theories of the urban housing market as well as the links between an urban hous-

Segmentation, Adjustment and Disequilibrium

55

ing market and other urban sectors should be constructed. For example, the existence and the different dynamic patterns of urban housing sub-markets suggest that disaggregate urban house price indices at sub-market level should be constructed. The disequilibrium property inherent to an urban housing sub-market model suggests that any successful economic analysis of an urban housing market should take disequilibrium issues into consideration. Failure to do so may result in a biased understanding of the urban housing market and its processes. The sub-market approach also provides important potential insights into the impacts of urban policy or housing policy. Many urban or housing policies are aimed at having localised impacts; for example, neighbourhood urban regeneration programmes are aimed at improving the neighbourhood economy, which may have impacts on the housing sub-market in which the neighbourhood is located. The urban housing sub-market model suggests that these impacts can then be transmitted to the rest of the market.

Notes (1) The same outcome would occur in a multiple bidding auction where the purchaser’s initial offer price exceeds all potential competitors. (2) Search burdens are exacerbated by the slow adjustment of housing supply in response to any demand changes. Housing supply comes from three sources: existing dwellings for sale, newly modiﬁed dwellings for sale in response to market demand and newly constructed housing units. The supply from any of these sources will not be able to respond to demand changes instantaneously. Slow turnover in the second-hand housing stock (partially related to high transaction costs) together with lagged new housing completions from either modiﬁcation or new construction imply that the demand can change more rapidly than supply (Wood & Maclennan 1982; Rothenberg et al. 1991). (3) A similar deﬁnition can be found, for example, in Grigsby et al. (1987), while Allen et al. (1995) and Bajic (1985) deﬁne housing sub-markets in relation to different dwelling types and dwelling structures.

4 Transactions Costs and Housing Markets John M. Quigley Introduction The durability, ﬁxity, and heterogeneity of dwellings imply that transactions costs are signiﬁcant in the housing market. Certainly, relative to ﬁnancial markets and to the markets for most consumer goods, housing transactions require costly search to uncover the prices and attributes of commodities. Transactions may require complex negotiation with ﬁnancial institutions as well as bargaining among housing market participants. Moreover, active choice in the housing market is infrequent, so participants may ﬁnd housing market options and choices more uncertain than those in other markets. In particular, bargaining itself may be costly to households given its infrequent occurrence. Duncan Maclennan’s well-known 1982 book, a treatise and a widely used text (Maclennan 1982), recognises the importance of these features of the housing market. Indeed, the expanded framework offered in Chapter 3 of that book stands in stark contrast to that offered in conventional economic texts of the time. It is fair to say that one of the enduring contributions of the book is its emphasis on dynamic processes in the market for housing and the need to recognise these processes both in applied research and in economic policy. This paper provides a review of the most important sources of transactions costs in the housing market and an assessment of the magnitude of these deviations from the simple model of frictionless competition by fully informed actors. In the second section, below, I explore a taxonomy of costs associated with transactions in the housing market. These transactions costs are, of course, speciﬁc to the market institutions of Britain and North America, and most of the quantitative evidence presented on costs comes from the US. The third

Transactions Costs and Housing Markets

57

section presents a simple theoretical exercise to investigate the magnitude of transactions costs relative to housing consumption measured in terms of rent or housing values. The ﬁnal section considers some selected implications of the analysis.

A taxonomy As emphasised by Maclennan (1982, pp. 59–75), ‘the’ housing market is really not a single neoclassical exchange market, but is rather a set of overlapping sub-markets differentiated by tenure, location, size and quality. There are markets in which the capital good, housing, is exchanged, and there are other markets in which the consumption good, housing services, is exchanged. The impediments that inhibit the adjustment of the market to a neoclassical equilibrium can be summarised in ﬁve broad categories: search costs, legal and administrative costs, adjustment costs, ﬁnancial costs, and the costs of uncertainty. Each can be expressed in terms of identiﬁable components. Some of these components are quite hard to measure and quantify, but are nevertheless signiﬁcant.

Search costs The heterogeneity of housing and its unique spatial component suggest that it is costly to identify available dwellings. The essence of housing market choice is the trade-off among size, quality, and locational considerations, together with price. Identifying and evaluating these attributes typically involves the physical inspection of dwellings at disparate locations. Some of these inspections can be made simply by driving past a property, and others can be made by examining photographs or newspaper advertisements. These are the inspections that lead to the elimination of properties from consideration. Any property that is a ‘serious’ candidate for choice will no doubt be the object of physical inspection by a potential renter or purchaser. There is little survey evidence on the number of dwellings inspected by households before making residential choices. Clark (1993) reports on studies of Canadian and US housing markets that document that a third of all homebuyers and half of all renters consider only one alternative. Clark’s own research (1982b) found that recent homebuyers in Los Angeles searched for less than a month, within an area of about a three-mile radius, and looked at 15 houses. Survey evidence from Glasgow suggested that renters spent between 7 and 19 days searching and that a large fraction accepted the ﬁrst available unit (Wood & Maclennan 1982).

58

Housing Economics and Public Policy

In addition, some information is available on the time spent searching by low-income renters. This information, the average number of days spent searching for rental accommodations, was gathered as a part of the Experimental Housing Allowance Program (EHAP) in the US (Weinberg et al. 1981). In the two markets in which the experiment was conducted, the median household spent 61 days searching (in Pittsburgh), with a standard deviation of 19 days, and 37 days searching (in Phoenix), with a standard deviation of 18 days. If households spent merely ﬁve hours a week searching during these time intervals, they would have devoted the equivalent of between one half and one full work-week to searching for housing. This is a substantial expenditure of effort. For renters, there are rather limited opportunities to employ brokers to reduce search costs. For home purchasers, whose other transactions costs are higher (see below), there is every reason to expect that the time devoted to search will be greater. However, for institutional reasons, there is more scope for the substitution of the services of brokers or other middlemen to assist in the search process. By custom in the US, brokerage commissions are split between buyer and seller brokers. On a given transaction, the buyer’s broker commission may be 3% of the selling price of a dwelling. DiPasquale & Wheaton (1996) assert that aggregate realtor (or estate agency) fees for housing transactions range from 3% to 6% in the US. A recent paper by Wood (1996) documents the regulated real estate commissions in Australian states. Average commissions range from 2% to over 8%, and commissions are proportionately smaller on more expensive houses. Technology has already begun to reduce some of these costs for homeowners. Online services (see, for example, homegain.com) provide richer opportunities to observe the qualitative and quantitative aspects of housing alternatives, at least in comparison with multiple listing services. It may be possible to ‘view’ a large fraction of available housing alternatives using web-based technology in the near future. This has the potential to reduce buyers’ search costs, at least in markets characterised by multiple listing services rather than a tradition of exclusive listings by individual brokers. Despite this potential, search costs in the housing market will remain high. Technology may help to eliminate inappropriate alternatives from consideration, but ultimately serious contenders will require physical inspection.

Legal and administrative costs Legal and administrative costs are considerably different for consumers who

Transactions Costs and Housing Markets

59

choose to be renters rather than owners. Rental contracts may specify the payment of security deposits, key fees, and other costs. These may amount to a couple of months’ rent, but under typical contracts (Jaffe 1996), these fees are returned (sometimes with interest) at the time of lease termination. However, these fees may contribute to a cash ﬂow problem, especially for low-income renters. For home purchasers, the legal and administrative fees due at the time the contract is executed may be far larger. In many jurisdictions, ad valorem taxes are levied at the time of title transfer. These transaction taxes are widespread in continental Europe. Stamp duties, a form of ad valorem transfer fee, are levied in Australia as well as the UK and in many jurisdictions in the US. In some jurisdictions in the US and the UK, lawyers are present at conveyance, and substantial legal fees are incurred. Recording and conveyance fees are levied by local governments. In the US, most lenders require a new title search at the time a transaction is completed, often involving substantial legal costs. Finally, there may be costs associated with opening or transferring accounts for public utilities and local services. Together these transactions costs for home purchase may reach several per cent of the value of the house. DiPasquale & Wheaton (1996) estimate that closing costs are 1–3% of the purchase price. Work by Chambers & Simonson (1989) suggested that the aggregate transactions costs of homeownership amount to 6–10% of house value. Cunningham & Hendershott (1984) argue that the value is roughly 12% of house value. Malatesta & Hess (1986) report survey evidence: average homeowner transactions costs were 12% of house values based upon a sample of 100 movers. These costs are quite substantial.

Adjustment costs The adjustment costs of moving include both out-of-pocket and psychic costs. Out-of-pocket costs include the costs of moving possessions and, perhaps, the value of furnishings rendered unusable after the move. For lowincome renters, estimates are available from EHAP. For Pittsburgh households, median out-of-pocket moving costs were 12% of one month’s income; for Phoenix households, the median was 3% of a month’s income (Weinberg et al. 1981, Table 1; and Friedman & Weinberg 1981, Table 1).

60

Housing Economics and Public Policy

These estimates seem quite low. However, they refer to the intra-metropolitan moving costs of low-income renters. Presumably, short distance moves are cheaper than long distance moves, and poor households have fewer possessions. In contrast, it is reported by Allied Moving Company that the average cost of an interstate household move in the US was $9000 in 1998, roughly two and a half times monthly income. The psychic costs of moving to a different residence are much harder to ascertain. Presumably, these psychic costs are larger for long distance moves than for short distance moves. In principle, the psychic and transactions costs are revealed by the maximum amount that a household of static socio-economic characteristics is willing to pay to continue residence in its current dwelling. An early paper by Dynarski (1986) drew attention to this concept. Papers by Venti & Wise (1984) and by Bartik et al. (1992) provide estimates of this willingness to pay. Both estimates are derived from EHAP data. Thus, they are likely to be underestimates of the willingness-to-pay of middle-income households, especially owners. The estimates are quite large indeed. The Venti & Wise methodology suggests that the average household in the EHAP sample would require a 14% increase in utility to make moving to another location as attractive as staying. The income equivalent is about $60 a month for these low-income households – or roughly 50% of their monthly rental repayments. The methodology employed by Bartik et al. suggests that the average low-income household in Phoenix (Pittsburgh) has moving costs, out-of-pocket plus psychic costs, of 25% (14%) of income. Households with longer periods of tenure have total moving costs that are larger still. Households whose heads are ten years older than the average have moving costs, out-of-pocket plus psychic costs, that are estimated to be 34% of income in Phoenix and 24% in Pittsburgh. The estimates of moving costs are yet higher for minority households, presumably reﬂecting their higher search costs.

Expectations and uncertainties Household uncertainty and expectations about the future can increase the transactions costs of moving. Expectations of declines in interest rates make homeowner mobility more expensive today if there are any costs to contracting (Quigley 2002), even with variable rate mortgages. Expectations of falling house prices make homeowner mobility cheaper today (Chan 2001). Expectations about tax changes affect the ex ante user cost of capital, and this affects the mobility decisions of households and their homeownership propensities (Rosen et al. 1984). There are many similar instances in which

Transactions Costs and Housing Markets

61

uncertainty and expectations affect ex ante transactions costs and ultimately residential mobility. Quantitative estimates are hard to come by.

Financing costs Beyond the transactions costs of negotiating and recording contracts, securing title, and so forth, there may be purely ﬁnancial costs associated with housing market transactions, and these costs may be quite large for some homeowners. For owner-occupants with ﬁxed rate mortgage contracts, increases in market interest rates may increase the value of the mortgage contract itself. When rates increase, the right to make monthly payments at the contract interest rate may have quite a large present discounted value. Unless mortgages are completely assumable, this value would be completely dissipated by moving to another residence. This factor alone can greatly increase the transaction costs of changing residences. It has been reported that the incidence of homeowner mobility declines substantially when interest rates rise (Quigley 1987, 2002), and homeowners are more likely to invest in renovations and home improvements as an alternative to incurring large transactions costs (Potepan 1989). Declines in house prices increase the costs of mobility that would force households to realise capital losses on their homes (Stein 1993; Chan 2001). Transactions in a declining market are less numerous due to loss aversion by sellers (Genesove & Mayer 2001). This also drives up search costs. Other institutional aspects of the market for mortgage ﬁnance may increase the transactions costs of home purchase for some households or may shut households out of the market for owner-occupied housing entirely. For example, the inability to borrow against human capital interacts with the mechanics of the level payment mortgage to make it impossible for some younger households to enter the homeownership market at all, even though the actuarial risk of default is low (see the chapter by Gyourko in this volume).

A simple model of transactions costs As suggested in the previous section, the extent of transactions costs in the housing market is large, and the sources of transactions costs are myriad. There is some survey evidence on the magnitude of these costs, at least for some households. But this evidence is noticeably incomplete. A full enumeration of the magnitude of these costs, especially as they affect different kinds of households, would be a formidable undertaking.

62

Housing Economics and Public Policy

A simple model may help provide a benchmark for the magnitudes involved. For illustration, suppose the utility function of consumers is Cobb-Douglas in housing (H) and other goods (X). The unit price of housing is PH; the price of the numeraire is set to unity. Thus, U = Hα X1–α = Hα (Y – PHH)1–α

(4.1)

where Y is household income and α is a parameter. This implies a demand function for housing PHH = R = αY

(4.2)

where, in equilibrium, the household spends a ﬁxed fraction of its income on rent (R). Now consider a household observed to be consuming HO or paying RO in rent with income YO. (In general this rent payment need not be the equilibrium expenditure, αYO.) The income (Y*) required to make the household as well-off as if it incurred the transactions costs associated with moving to consume its preferred housing bundle H*, is the solution to (4.3) In terms of rent,

(4.4)

Thus, the income equivalent is

(4.5)

Clearly, if the household is consuming its desired equilibrium level of housing services (i.e., if RO = αYO), the right-hand side of (4.5) is zero. If the household is not consuming its desired equilibrium level of housing services, this is because the transactions costs in the housing market for this household are larger than the income equivalent of the gain in utility from changing dwellings. Thus, the income equivalent, , is a lower bound estimate of the transactions costs incurred in the housing market. For households with transactions costs lower than , housing market adjustment will occur. For households who do not make market adjustments, must be less than the costs they would incur by being active participants in the market.

Transactions Costs and Housing Markets

63

If the demand function for housing is known, say PH = D(H,Y), but not the utility function, then the difference in consumer surplus provides an approximation to the income equivalent (4.6) For example, if the demand curve for housing services is log-linear, log H = log Z + α log P + β log Y

(4.7)

then substitution into (4.6) yields (4.8) or, in terms of rent (4.9)

Again, is a lower bound estimate of the transactions costs associated with housing market transactions. Of course, the model sketched out above is deceptively simple. In particular, it abstracts from issues of dynamic adjustments, as households’ demands for housing change over time and over the life cycle. Any calculations based on it are certainly very crude.1 Housing is a vector of attributes, not a scalar, and these attributes are priced jointly in some hedonic framework. The demand for these attributes varies for households of different characteristics (see Malpezzi’s chapter in this volume). Estimation of such a model would not be straightforward. But the essential point, that discounted transactions costs in the housing market are at least as large as the discounted utility gains forgone by consumers’ inertia, is an alternative to cumbersome surveys in quantifying these costs.

Some policy implications The magnitude of transactions costs in the housing market raises at least two issues: ﬁrst, the effect of large transactions costs on the rest of the economy; and second, the set of policies that might reduce these costs. It has been argued that the transactions costs inherent in home ownership affect the macro-economy by leading to increases in unemployment rates.

64

Housing Economics and Public Policy

In particular, A. J. Oswald (1997a, 1999) has argued that homeowners are less mobile than renters and are less willing to move to areas of job growth when they become unemployed. This is true, he argues, precisely because the transactions costs of the housing market are large for owners, relative to renters. Apparently, the only evidence offered in support of the proposition is a series of cross-tabulations and bivariate relationships. Across OECD countries, across US states and European regions, the simple correlation between homeownership rates and unemployment is positive. Recently, the proposition has received some theoretical support. More precisely, logically consistent models have been developed which possess this property. For example, Haavio & Kauppi (2001) have developed an intertemporal multi-region model with stochastic business cycles in which owner-occupied and rental housing markets are imbedded. With permanent boom-and-bust towns, or with random uncorrelated cycles, owners suffer no capital losses and owner occupation is just as efﬁcient as renting (in which case all capital effects are borne by absentee landlords). However, with any persistence in business cycles, some homeowners suffer repeated capital losses. With exogenous borrowing constraints, these households are unable to move to booming regions where employment opportunities are expanding. Stochastic capital losses represent a kind of transaction cost that arises from debt constraints, not risk aversion. The general hypothesis that homeownership rates and unemployment rates are correlated has been investigated more systematically for the United States by Green & Hendershott (2001a). The authors investigate state level unemployment and homeownership rates. They abstract from ﬁxed effects by analysing ﬁrst differences between 1970 and 1990. When the bivariate relationship is tested, weighting by population, any simple correlation vanishes (in fact, the estimated coefﬁcient is negative). In a series of regressions using age-speciﬁc measures of homeownership and unemployment, the coefﬁcients of homeownership are again generally insigniﬁcant. Green & Hendershott (2001b) compare the magnitude of the estimated effects for household heads and for secondary workers. They do ﬁnd differences, leading them to conclude that ‘tenure seems to inﬂuence labour decisions of those for whom the transaction cost of owning is large relative to the cost of not ﬁnding a job immediately’. Of course, households are not randomly assigned to the category of owner or renter, and this has important implications for the interpretation of these results. As documented earlier, the transactions costs of homeownership are much larger than the transactions costs of renting, and many of these are

Transactions Costs and Housing Markets

65

ﬁxed costs incurred at the time of the move. Presumably, households take this factor into account in choosing between homeownership and rental status. Thus, the observed and unobserved factors that cause households to expect to be long-term residents affect the probability that an ‘otherwise identical’ household will choose home ownership over rental status. This selectivity bias surely leads to an overestimate of the effect of the transactions costs of homeownership on unemployment in the results reported by Green & Hendershott.2 Thus, it does not appear that the transactions costs associated with the housing market are of much consequence to the functioning of the labour market. At least, there is no credible evidence that the institution of homeownership ‘causes’ higher unemployment levels in the economy. This does not mean, however, that there is no cause for concern about the extent of transactions costs in the housing market. Many of these costs are incurred simply as a waste of resources (e.g., time spent searching). Others represent fees to market intermediaries (and thus appear as national income), but the resources may be better employed elsewhere. Government can assist in improving this allocation of resources. It can play an active role in facilitating ﬂows of information and public goods. Government can also provide a forum for the development of standards that can make transactions simpler. The well-functioning homeownership market in regions where solicitors and lawyers are not active participants suggests that standardisation can reduce legal fees. Online property records can be extremely useful in many planning and ﬁnancial activities of local government. However, these online records would also provide an external beneﬁt to housing market participants by making title search instantaneous. Higher levels of government can efﬁciently ﬁnance this external beneﬁt. Property referrals for renters could economise on search costs for demanders and vacancy costs for suppliers. Some of these cost reductions will arise naturally as cheap information technology proliferates. But if these technologies are compartmentalised in parts of the market, through exclusive listings of various kinds, the advantage of cheaper information will not be fully realised. Governments and professional organisations can help in the production of these public goods.

Acknowledgements This chapter beneﬁted from the comments of Christine Whitehead, Geoff Meen, and Gavin Wood, and from the assistance of Tracy Gordon.

66

Housing Economics and Public Policy

Notes (1) One simple example: consider a ﬁve-year time horizon, a household of $40 000 income, and a structure of preferences in which households spend 25% of their income on housing in equilibrium. If we observed a household spending 30% of its income on housing, we could conclude (from equation 4.6, above) that the transactions costs in the housing market for that household were at least 10% of its monthly housing expenditures. (2) This is clearly recognised by the authors. Indeed, a more recent analysis by the authors (Green & Hendershott 2001b) attempts explicitly to address this selectivity issue.

5 Hedonic Pricing Models: a Selective and Applied Review Stephen Malpezzi

Introduction Twenty-ﬁve years ago, a few years before his Housing Economics, Professor Duncan Maclennan published a seminal paper entitled ‘Some Thoughts on the Nature and Purpose of House Price Studies’. The two publications were well-timed: hedonic price modelling was in the process of moving from a cutting-edge empirical curiosity to a standard method of price index construction that has been used in literally thousands of studies since.1 In those two publications and in Maclennan’s subsequent work, several important themes recur:

•

the need to put hedonic models on a ﬁrm theoretical footing, including, but not limited to, consideration of the consequences of disequilibrium;

•

whether common speciﬁcations are complete, or reasonably so; including, but not limited to, questions of omitted variables, functional form, and the proper deﬁnition of a market; and the implications for our work of inevitably imperfect speciﬁcations; and

•

that the design of the pricing model should ﬁt the purpose at hand.

In this chapter I selectively review the hedonic price literature (and, brieﬂy, some other related models) with a focus on these questions. Theory will be discussed; but my orientation is more towards the applied economist estimating these models, rather than specialised theorists. In particular, several other recent surveys such as Follain & Jimenez (1985a) and Sheppard (1999) have discussed certain theoretical and econometric issues which I will

68

Housing Economics and Public Policy

discuss only brieﬂy below; readers are referred to those excellent reviews, which I aim to complement rather than compete with.

What is a hedonic price index? The method of hedonic equations is one way that expenditures on housing can be decomposed into measurable prices and quantities, so that rents for different dwellings or for identical dwellings in different places can be predicted and compared. At its simplest, a hedonic equation is a regression of expenditures (rents or values) on housing characteristics. The independent variables represent the individual characteristics of the dwelling, and the regression coefﬁcients may be transformed into estimates of the implicit prices of these characteristics.

The fundamental hedonic equation Hedonic regressions are basically regressions of rent or house value against characteristics of the unit that determine that rent or value. The hedonic regression assumes that one knows the determinants of a unit’s rent: R = f (S, N, L, C, T)

(5.1)

where R = rent (substitute V, value, if estimating hedonic price indices for, say, homeowners using sales data); S = structural characteristics; N = neighbourhood characteristics; L = location within the market; C = contract conditions or characteristics, such as whether utilities are included in rent; and T = the time rent or value is observed. In this chapter, I will refer to a hedonic model more or less along these lines as a ‘single equation’ model or the ‘ﬁrst stage’ of a ‘two-stage’ model. Twostage models attempt to go beyond the initial estimation of a hedonic price

69

Hedonic Pricing Models: a Selective and Applied Review

surface, and in the second stage recover structural supply and demand parameters for individual housing characteristics. Collapsing the vectors S, N, L and C into a larger vector X for the moment purely for notational convenience, and adopting a common (but sometimes criticised, see below) semi-logarithmic functional form, (5.1) can be re-written compactly as: R = exβε

(5.2)

so that ln R = Xβ + ε

(5.3)

and we estimate: ln R = Xb + e

(5.4)

where β and ε are of course the unknown true parameters, and b and e are actual estimates. Now, by properties of logarithms, the predicted rent of a unit can be computed as R = exb; the price of an individual attribute, X1, at a given level of X1, given the level of the other m-1 attributes, Xi≠1, can be calculated in dollars or pounds as: P = exb

(5.5)

Notice that with such a logarithmic speciﬁcation, the dollars or pound price of X1, or any other single characteristic, varies with the level of X1, as well as with the level of other Xi. Prices are non-linear, an important point to which I return below.

‘Second-stage’ hedonic models: recovering structural parameters Much of the hedonic literature focuses on the basic hedonic relationship discussed in the preceding paragraphs. However, as papers discussed below by Rosen and others make clear, the hedonic equation discussed above is a reduced form. Under certain maintained hypotheses hedonic equations also admit of a structural interpretation: for example, if the supply of each and all characteristics is perfectly elastic, hedonic coefﬁcients reveal demand for characteristics. But in most real-world contexts such a stringent maintained

70

Housing Economics and Public Policy

hypothesis is untenable. A number of papers attempt to recover structural parameters of demand and supply; or at least the demand for characteristics. Speciﬁcally, because dollar or pound prices vary within a sample, if the level of characteristics also varies, one can make use of this variation to estimate price elasticities for individual coefﬁcients. For example, one speciﬁcation that has been used in a number of papers is to presume a linear demand model in the second stage, after a ﬁrst-stage logarithmic hedonic. It is not uncommon to place prices of each characteristic on the left-hand side, i.e., to assume an inverse demand relation, and to then estimate an equation for each characteristic, of the form: (5.6)

where P1i = eXb as before. Note that the money price of each of the m characteristics will vary from one observation to the next (will vary with i) because of the property of joint determination of prices discussed above. Armed with this variation in price, demand estimation can proceed. Papers such as Follain & Jimenez (1985b) and Witte, Sumka & Erekson (1979) present estimates of the demand for housing characteristics from such models. Vectors D and S represent exogenous demand and supply shifters, such as income, or input costs (typically land). Often supply is assumed elastic, so that only demand shifters, like household income or family size, that are more readily available, are included. Thus, a prototypical dataset for the ﬁrst stage would include household level data on some dependent variable like rent or sales price; and on the characteristics S, N, L and C. The prototypical dataset for the second stage would add information on household/unit level demand and supply shifters. Sometimes neighbourhood-level data are appended to household level data.2 A point now well understood by most experienced practitioners, but only occasionally discussed in the literature, is just how central a role functional form plays in the set-up for most two-stage demand for characteristics models. Consider that if the ﬁrst-stage hedonic regression were linear, there would be no variation in characteristic prices within the sample, and hence no second-stage system to estimate. In fact, it is the difference between hedonic functional forms, and second-stage demand functional forms, that

Hedonic Pricing Models: a Selective and Applied Review

71

makes such systems potentially estimable. This point was made clearly by Nelson’s (1982a) insightful critique of Witte, Sumka & Erekson’s (1979) otherwise exemplary early study. Witte et al. (WSE) estimated logarithmic hedonic and logarithmic demand functions; Nelson showed that these were not in fact estimable, and Nelson suggests that the fact that WSE did obtain numerical estimates was, paradoxically, due to rounding error. On the other hand, models that involve (say) logarithmic ﬁrst-stage hedonic equations and linear second stage demand equations may be estimable (though subject to remaining problems discussed below).

Repeat sales models In this chapter, the focus is on hedonic pricing models. In order to place this class of models in context, consider brieﬂy another pricing model, namely repeat sales models, which are in fact related to hedonic price indices in a certain way.3 Repeat sales indices are estimated by analysing data where all units have sold at least twice. Such data allow us to annualise the percentage growth in sales prices over time.4 These are time series indices in their pure form. They do not provide information on the value of individual house characteristics or on price levels. They have the advantage of being based on actual transactions prices, and in principle allow us to sidestep the problem of omitted variable bias. One way to understand the key features of the repeat sales index is to start by reconsidering the hedonic model. Consider a simple semi-log hedonic equation: ln P = Xβ + β1T1 + β2T2 + β3T3 + β4T4

(5.7)

where P is the value or rent for the unit, and where the vector X includes all the relevant characters, including a constant term; and the time dummies Ti represent periods that follow the initial base case period.5 The vector X represents a list of housing and neighbourhood characteristics that would enter a hedonic equation. The vector T is a series of dummy variables representing the time periods under consideration. These could be months, quarters, or years, depending upon the type of data at hand. Consider a house ‘A’ that sells in periods 2 and 4 (period 0 is the base year). In period 2:

72

Housing Economics and Public Policy

ln PA2 = Xβ + β1T1 + β2T2 + β3T3 + β4T4

(5.8)

= Xβ + β2T2 since T1, T3, and T4 = 0. And of course, by similar reasoning, in period 4: ln PA4 = Xβ + β4T4

(5.9)

Then, by subtraction: ln PA4 – ln PA2 = Xβ + β4T4 – Xβ – β2T2

(5.10)

= β4T4 – β2T2 This is for a representative housing unit that sells twice. Given a sample of such units, we want, in effect, the ‘average’ β4 and β2. Recall that regression is, in effect, estimating a series of conditional means. Clearly, by subtraction the characteristics vector drops out, as do the dummy variables for periods in which no transaction takes place. Green & Malpezzi (2001) illustrate with sample data. Another possible reﬁnement is to consider the fact that the variance of these housing prices will generally increase over time. In today’s econometric parlance, such prices are not stationary. Case & Shiller (1987) suggest a reﬁnement to the Bailey, Muth and Nourse model to mitigate such problems. The model just described is used as a ﬁrst stage, and the residuals from this ﬁrst stage model are used to construct weights that can be used to correct for heteroskedasticity using generalised least squares. Repeat sales indices are currently much discussed in the literature because they have a number of advantages. First, no information is required on the characteristics of the unit (other than that an individual unit has not signiﬁcantly changed its characteristics between sales). Second, the method can be used on datasets that are potentially widely available, at least in the US, and collected in a timely manner, with great geographic detail, but which do not have detailed housing characteristics. For example, Case and Shiller’s original work used data collected by the Society of Real Estate Appraisers. Much of the current US research in this area has been undertaken by Fannie Mae and Freddie Mac, which have the advantage of large datasets with price data from a huge number of transactions nationwide. The repeat sales method has a number of shortcomings as well.6 First, while raw data have been widely available in the US, data are often harder to come

Hedonic Pricing Models: a Selective and Applied Review

73

by in other countries, including Britain. Second, even at its best, the method only yields estimates of price changes. No information on price levels, or place-to-place price index, is derivable from the repeat sales method. Of course, the repeat sales method can be combined with some other method; i.e., to update earlier estimates of price levels constructed using some other method. Also, because only a few units transact twice over a given time period, the repeat sales method utilises only a fraction of potential information on the housing market. Other potential issues with repeat sales include the following. Units that transact frequently may be systematically different from units representative of the stock as a whole. How big a problem this selection bias is depends partly on the purpose of the index. It certainly would be less of a problem if the purpose of the index were to track the prices of units available on the market. The method also implicitly assumes that there is no change in the quality or quantity of housing services produced by the unit between periods. Of course, this assumption is always violated to some degree. Those who construct these indices spend a lot of time weeding out units that have been upgraded using, for example, collateral data on building permits, or the limited structural information that may exist in the dataset in use. The method also assumes that the coefﬁcients on the underlying hedonic model remain constant: this is what allows the house characteristics to drop out of the model. But this assumption may also be questioned. For example, as families have become smaller, so too has the value of bedrooms, holding all else equal. Thus the hedonic coefﬁcient for bedrooms in 1990 was almost certainly different from the coefﬁcient in 1960, regardless of the particular market.7

The roots of hedonic price models In essence, the hedonic relation arises because of heterogeneity. The model postulates a market containing a heterogeneous housing stock, which can only be modiﬁed at some cost, and heterogeneous consumers, some of whom put different valuations on a given bundle of characteristics (‘house’) than others. A full history of hedonic pricing models would be another paper, at least, but it is worthwhile to identify the roots of this approach, and point out some of the classics of the literature here. Two oft-cited classic papers are those by Kelvin Lancaster (1966) and Sherwin Rosen (1974).8 Focusing on the demand side of the market, Lancaster developed a sophisticated branch of microeconomic theory in which utility

74

Housing Economics and Public Policy

is generated, not by goods per se, but by characteristics of the goods. The applicability to housing is direct and obvious. I’m happy to be home, not so much to be in anything called a ‘house’, but to be in a warm dry place, with a quiet space for a comfortable chair, a functioning toilet or a hot bath should I require them, and some other rooms in the house to store stacks of papers or noisy children.9 Thus, many hedonic studies cite Lancaster’s work, and justiﬁably so, for providing microeconomic foundations for analysing utility-generating characteristics. Lancaster developed this theory using the tools of ‘activity analysis’, and did not limit his discussion to housing, but applied the concept to topics as diverse as ﬁnancial assets, the labour/leisure trade-off, and the demand for money. Perhaps Lancaster’s main contribution was to put centre stage the still-slippery question of what exactly is the ‘good’ housing, and how is it related to more fundamental characteristics? Rosen’s (1974) article is the other oft-cited classic reference. Like Lancaster, Rosen focuses on characteristics, but has less to say about their utility-bearing nature and more about how suppliers and consumers interact within a framework of bids and offers for characteristics. Furthermore, while he did not much discuss functional form explicitly, Rosen’s model naturally leads to a non-linear hedonic price structure. Many two-stage characteristic demand models, in particular, cite Rosen as their theoretical foundation, although Rosen had little to say about how the estimation of such structural parameters might be carried out. Of course, there were other early theoretical papers that contributed to the development of this literature. One might point to Sir John Hicks’ (1939, 1960) elaboration of a ‘composite commodity’, and to Fisher & Shell’s (1971) related ‘repackaging hypothesis’, anticipating other early work by Triplett (1974) among others. As regards the actual estimation of hedonic price models, the study most often cited as the pioneer was not a housing application at all, but a hedonic price index for automobiles developed by A. T. Court (1939). A later, but still early and inﬂuential automobile application was by Griliches (1961). Recently, Goodman (1998) and Colwell & Dilmore (1999) have ﬁlled in more of this history, telling us more about Court’s early work as well as even earlier efforts by Wallace (1926) and Haas (1922). I have seen less discussion of which two-stage study has pride of place; Witte et al. (1979) and Awan et al. (1982) certainly helped get the ball rolling.

Hedonic Pricing Models: a Selective and Applied Review

75

Conceptual issues in hedonic modelling In this section, I discuss brieﬂy some matters of underlying theoretical importance in hedonic modelling. The discussion is short because the issues have been so well discussed in two recent surveys, by Follain & Jimenez (1985a), and Sheppard (1999), as well as in several of the other papers cited in the next few paragraphs.

Two identiﬁcation problems The identiﬁcation problem – disentangling supply and demand, when faced with data only from their interaction – has bedevilled applied econometrics for years. In addition to the ‘usual’ identiﬁcation problem, two-stage hedonic analysis of the demand for characteristics faces an additional potential problem. The problem stems from the nonlinearity of the price structure. In garden-variety supply and demand models, individual consumers (and, often, suppliers) are price-takers, i.e., the price of the good is exogenous and the consumer chooses a quantity conditional on the reigning price. In non-linear hedonic models, whether a simple logarithmic model or a more ﬂexible form, prices and quantities are correlated by construction; in effect, consumers choose both a quantity of some characteristic and, implicitly, its price. The problem has been well analysed by Blomquist & Worley (1982) and Diamond & Smith (1985), and in the Follain & Jimenez (1985a) and Sheppard (1999) surveys; the reader is referred to their more discursive treatment. Sufﬁce it to say that a number of studies have tried to tackle the problems using well-established techniques like instrumental variables. The problem, as ever, is in ﬁnding good instruments.

Equilibrium or disequilibrium models? Another ubiquitous feature of housing markets is that their extremely costly adjustment processes make the usual simplifying assumptions that markets are in equilibrium when observed less tenable. In fact, the disequilibrium nature of the housing market is a recurring theme in Maclennan (1982) and in many of his other papers. What are the consequences of this disequilibrium, and what can be done about it? Several possible approaches to issues arising from disequilibrium are possible. In addition to the examples included here, see also the chapter by Yong Tu in this volume. One possible approach to the problem of disequilibrium is to estimate hedonic price functions using only observations in or near

76

Housing Economics and Public Policy

equilibrium. For example, a switching regression approach could be used, following generally the econometric methods described in, for example, Bowden (1978). Such approaches have been applied generally to the housing market, notably by Fair & Jaffee (1972); but their models are generally studies of the determinants of housing starts, not models of housing prices. An example of a later disequilibrium hedonic model is Anas & Eum (1984). Several difﬁculties must be overcome to implement such a switching regression model. First, it is necessary to specify the nature of the process that distinguishes equilibrium from disequilibrium observations; in a typical cross-section hedonic sample it is not always obvious how this would be done. Second, depending on the purpose of the hedonic index, successfully estimated equilibrium prices may or may not be what it is needed. For example, if one were constructing place-to-place price indices to study, say, the cost of living or to set appropriate levels of housing subsidies, presumably one would wish the index to reﬂect actual prices paid in the market, whether in or out of equilibrium. An alternative approach is to focus, not on the effects of disequilibrium on the construction of the price index, but rather on the amount of disequilibrium in a given sub-market or period, and the adjustment process back to equilibrium. Examples of studies that illustrate the principles include Abraham & Hendershott (1996), Dreiman & Follain (2000), and Malpezzi (1999), each of which studies prices over time. While different in important details, each of these studies essentially proceeds in three steps. First, estimate a time series price index.10 Second, ﬁnd some way of differentiating equilibrium prices; in these studies, prices were considered near equilibrium if price changes in the next period were near zero. With this subset of equilibrium prices, estimate their fundamental determinants, e.g. based on income, recent growth, supply conditions, etc. Now for each period we have the price and index actually realised, and an estimate of the equilibrium price; thus we have an estimate of disequilibrium as well. The third and ﬁnal phase of these studies is to study the determinants of the disequilibrium, including in these time-series studies the nature of the time path back to equilibrium, once the market has been ‘shocked out’.

Speciﬁcation issues In this section, I discuss a number of practical issues such as which variables or functional form to use, and how to deﬁne the geography of a market. But ﬁrst, one should note an overarching reason why decisions on these matters so often seem ad hoc. It is unfortunate, but the answer to ‘what does theory

Hedonic Pricing Models: a Selective and Applied Review

77

tell us about speciﬁcation of hedonic models?’ is, in brief, ‘not much’. Papers like Lancaster (1966) and Rosen (1974) elegantly present models of housing characteristics without having much to say about just what those characteristics are, or how exactly they are related to price.

Choice of dependent variable Firstly, the choice of dependent variable is about choosing rent, or value, of the housing unit. Confusion sometimes reigns because of sloppy terminology, especially among real estate professionals, but also sometimes among housing economists. The term ‘housing price’ is often used loosely as a synonym for ‘housing value’, when of course it is a true ‘price’ only under special conditions. But the usage is so entrenched by now that even housing economists generally rely on context to keep the meaning clear. Of course it is well known that house rents and house values are related, though only in special cases proportionately so; papers such as Ambrose and Nourse (1993) and Phillips (1988) explore systematic variations in the relationship, or ‘capitalisation rate’. Papers that focus on rents must wrestle with problems stemming from the fact that different units have different lease terms or contract conditions. One notable example is the inclusion, or exclusion, of utility payments in rent. One common procedure is to obtain data on such utility charges for units where they are not included, and to add these charges to contract rent to ‘gross it up’, so that rents in the sample are for comparable services. Lump sum payments (deposits or ‘key money’) can be annualised with an assumed capitalisation rate, and added to rent as in Malpezzi’s (1998) study of key money paid for Cairo rent controlled units. Another is to use contract rent as the dependent variable, but to add dummy variables to the right-hand side to indicate units with various utilities included in rent, so that the estimated coefﬁcients ‘dummy out’ the price of utilities, leaving a rental index net of utilities. When estimating a hedonic regression on values, several other measurement issues emerge. A number of studies use owner or tenant estimates of the value of the unit. This gives rise to concern over the accuracy of such selfreported appraisals. Several papers have examined the issue with US data, such as Kain & Quigley (1972b), Follain & Malpezzi (1981b), and Goodman & Ittner (1992). As yet I have found few non-US studies addressing the issue. Early studies such as Kain & Quigley and Follain & Malpezzi suggested that while the variances of owner assessments are high, biases are modest; given enough data, hedonic models based on owner assessments would be

78

Housing Economics and Public Policy

reasonably reliable. But Goodman and Ittner’s recent study ﬁnds larger biases, and suggests more caution. Recent sales ‘prices’ (house values from observed recent transactions) have some obvious advantages as dependent variables. Recent transactions data may present less potential bias, and greater potential precision, than occupants’ or owners’ self-assessments. But recent sales are not necessarily a random draw from the total housing stock. If the purpose is to index the market of available units, this may not be of great concern, but if the purpose is to index the total stock, we must concern ourselves with possible selection bias. Several papers such as Gatzlaff & Haurin (1997) have tested the presence of such biases. Test statistics often reject the null, but so far most studies have found the magnitude of the bias to be modest. Some datasets, like the American Housing Survey, truncate housing values; values over $300 000 are reported simply as ‘over $300 000’. Such dependent variable truncation can cause signiﬁcant bias in results. Maddala (1983) presents some econometric techniques that attempt to attenuate the effects of such bias, when truncation is an issue.

Selection of independent variables There are literally hundreds of potential housing characteristics that could be included on the right-hand side. Butler (1982) and Ozanne and Malpezzi (1985) show that, unfortunately, coefﬁcient estimates are not robust with respect to omitted variables. But interestingly, the same correlation between omitted and included variables that biases individual coefﬁcient estimates can and often does help improve prediction from a ‘sparse’ model. This suggests that hedonic applications that rely on overall predictions – like placeto-place price indices, or cost-beneﬁt analysis of housing subsidies – can proceed apace, even while papers that rely on interpretation of individual coefﬁcients must be interpreted more cautiously. While theory is not much of a guide, experience from many studies suggests that, whatever the purpose, a full dataset would include the following:

•

rooms, in the aggregate, and by type (bedrooms, bathrooms, etc.)

•

ﬂoor area of the unit

•

structure type (single family, attached or detached, if multi-family the number of units in the structure, number of ﬂoors)

Hedonic Pricing Models: a Selective and Applied Review

79

•

type of heating and cooling systems

•

age of the unit

•

other structural features, such as presence of basements, ﬁreplaces, garages, etc.

•

major categories of structural materials, and quality of ﬁnish

•

neighbourhood variables, perhaps an overall neighbourhood rating, quality of schools, socio-economic characteristics of the neighbourhood

•

distance to the central business district, and perhaps to sub-centres of employment; access to shopping, schools and other important amenities

•

characteristics of the tenant that affect prices: length of tenure (especially for renters), whether utilities are included in rent; and possibly racial or ethnic characteristics (if these are hypothesised to affect the price per unit of housing services faced by the occupant)

•

date of data collection (especially if the data are collected over a period of months or years).

However, this list, while still incomplete, is also general. Hocking (1976), Amemiya (1980) and Leamer (1978) are among useful guides to the actual selection of variables.

Functional form in general There is no strong theoretical basis for choosing any speciﬁc functional form for a hedonic regression (see Halvorsen & Pollakowski (1981) and Rosen (1974)). Follain & Malpezzi (1980b), for example, tested a linear functional form as well as a log-linear (also known as semi-log) speciﬁcation. But they found the log-linear form had a number of advantages over the linear form, detailed below. The log-linear form is written: ln R = β0 + Sβ1 + Nβ2 + Lβ3 + Cβ4 + ε

(5.11)

where ln R is the natural log of imputed rent, S, N, L and C are structural, neighbourhood, locational, and contract characteristics of the dwelling,11

80

Housing Economics and Public Policy

and βi and ε are the hedonic regression coefﬁcients and error term, respectively. The log-linear form has ﬁve things to recommend it. First, the semi-log model allows for variation in the dollar value of a particular characteristic so that the price of one component depends in part on the house’s other characteristics. For example, with the linear model, the value added by a third bathroom to a one-bedroom house is the same as it adds to a ﬁve-bedroom house. This seems unlikely12. The semi-log model allows the value added to vary proportionally with the size and quality of the home. Second, the coefﬁcients of a semi-log model have a simple and appealing interpretation. The coefﬁcient can be interpreted as approximately the percentage change in the rent or value given a unit change in the independent variable. For example, if the coefﬁcient of a variable representing central air conditioning is 0.219, then adding it to a structure adds about 22% to its value or its rent. (Actually, the percentage interpretation is an approximation, and it is not necessarily accurate for dummy variables. Halvorsen & Palmquist (1980) show that a much better approximation of the percentage change is given by eb – 1, where b is the estimated coefﬁcient and e is the base of natural logarithms. So a better approximation is that central air conditioning will add exp (0.219) – 1 = 24%.) Third, the semi-log form often mitigates the common statistical problem known as heteroskedasticity, or changing variance of the error term. Fourth, semi-log models are computationally simple, and so well suited to examples. The one hazard endemic to the semi-log form is that the anti-log of the predicted log house price does not give an unbiased estimate of predicted price. This can, however, be ﬁxed with an adjustment (see Goldberger 1968). Last, it is possible to build speciﬁcation ﬂexibility into the right-hand side, using dummy (or indicator) variables, splines and the like (of which more shortly). This allows a fair amount of ﬂexibility in estimation, even with the semilog form. However, some authors have recommended more ﬂexible forms than the semi-log. One common ﬂexible form is the translog functional form, suggested by Christensen et al. (1973): (5.12) where ln R again represents the log of rent (value can be substituted), and there are m characteristics denoted X. Examples of the translog form can be found in Capozza et al. (1996, 1997).

Hedonic Pricing Models: a Selective and Applied Review

81

There is an even more general and ﬂexible class of functions, within which linear, logarithmic and translog functions are subsumed; these ﬂexible forms are carefully analysed by Box and Cox (1964), and applied to hedonic prices by Halvorsen and Pollakowski (1981): (5.13) Such a form is quite ﬂexible, with parameters θ and λ limiting the functional form.13 For example, when θ and λ are both 1 and γmn are all identically zero, the Box-Cox form becomes a simple linear model. When θ and λ approach zero and γmn are all identically zero, the Box-Cox form becomes a logarithmic model. When θ and λ approach zero and but some γmn are nonzero, the BoxCox form becomes the translogarithmic model. This is a good place to reiterate the special role functional form plays in two-stage structural models of characteristics demand and supply. I have already noted this important fact: it is functional form – indeed, differences in functional form between stages – that makes the system of demand (or supply and demand) functions potentially estimable. Thus it is particularly problematic that theory yields little guidance to the functional form of the hedonic relationship, and only tenuous guidance to the functional form for second-stage estimation of the demand for characteristics.

Functional form and independent variables If data permit it, judicious use of dummy or indicator variables for independent variables can be useful. For example, entering a variable for the number of total rooms in (say) a semi-logarithmic hedonic regression constrains the percentage increase in value from a one-unit addition to a 3-room unit to be the same as the percentage increase in value from a one-unit addition to a 6room unit. If degrees of freedom permit it, at least the most common values can be coded as dummy variables, imparting more ﬂexibility to the form. Malpezzi et al. (1980) provide additional details, e.g. how to code combinations of dummies and continuous variables. See also the classic review by Harold Watts (1964), and Suits (1984). The special topic of how to interpret dummy variables when the dependent variable is logarithmic is treated in Halvorsen & Palmquist (1980) and Kennedy (1981). Of course dummy or indicator variables are not the only method that can be used to incorporate ﬂexibility on the right-hand side. Continuous variables can be entered in quadratic (or cubic or even higher power) form; in fact, as

82

Housing Economics and Public Policy

much ﬂexibility as needed can be readily constructed using piecewise spline techniques (Suits et al. 1978).

Market and sub-market deﬁnition The deﬁnition and testing of sub-markets is an important recurring theme in Maclennan’s work. Housing markets are local and diverse, and hedonic price estimation requires careful consideration of sub-markets (see the chapter in this volume by Yong Tu). Sub-market assumptions in hedonic models can be roughly categorised as follows. The ﬁrst category comprises papers that deﬁne a market as an entire nation, or at least a large region, or perhaps a state. Linneman (1981) and Struyk (1980) fall into the category of national hedonic models, and Mills & Simenauer (1996) present a regional model. The second category, including much of my own work such as Malpezzi et al.(1980) and Follain & Malpezzi (1980a), adopts the metropolitan area as the unit of analysis. Metropolitan areas are usually thought of as labour markets, more or less, and it is certainly appealing to consider housing markets and labour markets as roughly coincident. The third category, including many of Maclennan’s own studies as well as papers such as Straszheim (1975), Gabriel (1984), Grigsby et al. (1987), Rothenberg et al. (1991), Maclennan & Tu (1996), and Bourassa et al. (1999), examines sub-markets below the metropolitan level. These may be segmented by location (central city/suburb), or by housing quality level, or by race or income level. Studies that obtain large datasets and test for the existence of sub-markets, usually by segmenting the sample and performing F-tests for equality of hedonic coefﬁcients across sub-samples, generally ﬁnd them: the F-tests usually reject the null. Ohta & Griliches (1975) suggest a more conservative method that focuses on changes in the standard error of the regression, in effect on how well the segmented model predicts.

Hedonic modelling: the current position Single-equation models, or structural models? Taken together, the problems discussed above – especially the identiﬁcation problems, imperfect speciﬁcations, and the general non-robustness of coefﬁcient estimates – suggest that reliable two-stage structural estimation

Hedonic Pricing Models: a Selective and Applied Review

83

of the demand for characteristics is difﬁcult. Qualitatively, that is the judgement we reached in the World Bank’s housing demand research project, after investing signiﬁcant resources attempting to develop characteristic demand models that would improve low-cost housing project design (Follain & Jimenez 1985b; Gross 1986; Mayo & Gross 1987). That does not mean that there is no hope for developing useful models somewhat along these lines. Models of aggregate housing demand work well enough, despite undoubted problems (see Mayo 1981; Olsen 1987; Whitehead 1999). King (1975) presents an ‘in-between’ model where housing is broken down into three categories – space, quality and location. Possibly further work along these lines would be fruitful.

Hedonic speciﬁcation: art or science? Generally, there is art as well as science in model speciﬁcation: choice of variables, functional form, and deﬁnition of sub-market. Whenever sample sizes are small, and especially if the application will involve some prediction out of sample (as with, say, pricing rent-controlled or subsidised units), it is often best to stick to a simple parsimonious speciﬁcation, possibly using the metropolitan area as market deﬁnition. But if samples are large and welldrawn, and especially if the focus of the hedonic model is a single metropolitan area, more ﬂexible forms, and more careful attention to the delineation of sub-markets, will generally pay off. It is somewhat surprising that the literature applying formal speciﬁcation tests, such as those of Hausman (1978), is modest, since speciﬁcation is such an issue in hedonic analysis. Burgess & Harmon (1982) is an interesting example that could be replicated further.

Examples of applications While there are many important and interesting theoretical issues related to hedonic models, some of which have been discussed above, our main interest ultimately is in understanding real-world housing markets, and hence in applications. Space precludes an exhaustive review, but here I mention some examples. I list a few representative applications mainly by topic; note also that while I know the US literature best and cite it heavily, hedonic models are now truly universally applied. As previously cited work by Maclennan (1982) and Ball (1973) make clear, there is of course a long-standing, large and growing literature focused on the United Kingdom, as well as the rest of

84

Housing Economics and Public Policy

Europe and of North America; but in fact hedonic models have been applied in every permanently inhabited region of the globe.14 One of the ﬁrst, and still most important, uses of hedonic models is to make general improvements in housing price indices, whether time series, placeto-place, or panel data price indices. Follain & Ozanne (1979), Chowhan & Prud’homme (2000), Englund, Quigley & Redfearn (1998), Follain & Malpezzi (1980b), Hoffman & Kurz (2002), Moulton (1995), Malpezzi, Chun & Green (1998) and Tiwari & Hasegawa (2000) are among many examples of studies that basically aim to improve the precision of housing price benchmarks. Some hedonic studies have been undertaken to construct special-purpose housing price indices, for example to improve the measurement of poverty thresholds (Short et al.1999). Hedonic prices have also been examined within cities. In addition to the submarket tests already discussed, many tests of the ‘standard urban model’ of Alonso, Muth and Mills have been carried out. The standard model predicts a generally declining pattern of prices with distance from the centre of the city. Competing models based on localised amenities, and models with multiple centres, have other predictions. Adair et al. (2000), Follain & Malpezzi (1981c), Mozolin (1994), Soderberg & Janssen (2001) are examples of studies that examine intra-urban variation in the price of housing using hedonic models. Perhaps unsurprisingly, results for the ‘standard model’ are mixed; while there are some broad tendencies for house prices to fall with distance from CBD, amenities and sub-centres generally play an important role as well (also, see the discussion and further references contained in the chapter by Gibb in this volume). Hedonic models have also been used to develop measures of environmental quality. One common approach is to examine whether house prices increase when near environmental ‘goods’, or fall when near ‘bads’. Examples of this literature include Cheshire & Sheppard (1995), Freeman (1979), Boyle & Kiel (2001), Des Rosiers & Theriault (1996), Din, Hoesli & Bender (2001) and Garrod & Willis (1992 a, b). Many other interesting studies have been undertaken that focus on interpretations of individual coefﬁcients. One must always be cautious in interpreting individual coefﬁcients in light of the speciﬁcation issues discussed above. With this caveat, a number of studies have examined racial, ethnic and socio-economic differences in housing prices. Kain & Quigley (1972a), Follain & Malpezzi (1981a), Chambers (1992), Galster (1992), Nelson (1982b) and Vandell (1995) are among many contributions to this strand of literature. Other studies have used hedonic age coefﬁcients to measure depreciation,

Hedonic Pricing Models: a Selective and Applied Review

85

such as Malpezzi et al. (1987), Clapp & Giaccotto (1998b), Goodman & Thibodeau (1995), and Shilling et al. (1991). Hedonic prices have been applied to market-rate units and then used to price subsidised or publicly provided units, in order to calculate the costs and beneﬁts of different housing subsidy programmes. Olsen & Barton (1983), Buchel & Hoesli (1995), De Borger (1986), Quigley (1982a), Satsangi (1991), Turner (1997), Gibb & MacKay (2001) and Willis & Cameron (1993) are representative examples. Closely related are studies that undertake regulatory cost beneﬁt, including rent control; for example, Olsen (1972), Malpezzi (1998) and Willis & Nicholson (1991). Another important use of hedonic models is the appraisal of individual housing units. Appraisers and other property market professionals increasingly use hedonic models. They can be used to improve professional practice of appraisers and chartered surveyors (Dubin 1998), or for undertaking mass appraisal for property taxation and other public purposes; see Berry & Bednarz (1975), Lusht (1976), and Pace & Gilley (1990). Hedonic models are also used to examine the capitalisation of a wide range of amenities, as well as costs. One of the earliest literatures along these lines developed to study whether differential local tax rates were capitalised into house prices, following the model of Tiebout (1956), as extended by Oates (1981). After several false starts, papers by Edel & Sclar (1974) and King (1977) clariﬁed the need to include measures of public services as well as taxes paid, and pointed out some important details of the correct functional form for such tests. Many subsequent studies have found such capitalisation, on both the beneﬁt and tax side; Zodrow (1983) provides a convenient review. Lastly, despite the problems discussed above, many studies have tried to recover demand parameters (and sometimes supply and demand parameters) for individual housing characteristics, or groups of characteristics. Studies here include those by Awan et al.(1982), Pasha & Butts (1996), Witte et al. (1979) and Kaufman & Quigley (1987).

Concluding thoughts A lot of cutting-edge work on conceptual issues relating to hedonics is still being done. Theoretical work continues apace to develop the foundations of hedonic models, and in particular to attempt to address some of the issues I have noted with two-stage structural models. In addition to the literature cited above, on this see, for example, Rouwendal (1992) and Epple (1987).

86

Housing Economics and Public Policy

An alternative way to think about hedonic models is a two-stage process of a different sort: samples used for hedonic estimation are not necessarily random draws from the population of houses, but are selected samples (especially when transactions-driven databases are used). Ermisch et al.(1996), Jud & Seaks (1994) and Clapp et al. (1991) are examples of studies addressing selection issues. In terms of functional form, one of the cutting-edge areas is to eschew parametric forms altogether. Semiparametric and nonparametric approaches can be found in Anglin & Ramazan (1996), Mason & Quigley (1996), Meese & Wallace (1991) and Pace (1993). Another approach is to use Bayesian restrictions on hedonic estimates, as outlined in Gilley & Pace (1995) and Knight et al. (1992). But perhaps one of the most exciting areas for extending hedonic models is making use of the spatial structure of the data, using the emerging technology of geographic information systems and spatial autocorrelation. Among other recent contributions in this area, see Can (1992), Dubin (1992), Pace & Gilley (1997), Basu & Thibodeau (1998), Gillen et al. (2001), and Thibodeau (2002). Thibodeau (2002), for example, ﬁnds a roughly 20% improvement in the ﬁt of hedonic models using these techniques. Especially in applications regarding mass appraisal, these techniques are extremely promising. However, there is also no end of applications that might not be thought of as cutting edge technically, that have not been done, but that are potentially terribly useful. Many of these are extensions of the studies listed in the previous section. Many housing programmes and policies have yet to be submitted to rigorous cost-beneﬁt analysis. Improving systems of mass appraisal remains important; for example, Russia is embarking on the development of a valuation system for all property in the entire country. While I have already cited individual hedonic studies from every continent, there is clearly scope for more and better international comparisons of housing prices. Recently there has been resurgence in cross-country comparisons, partly driven by the United Nations Centre for Human Settlements’ Housing and Urban Development Indicators project.15 Angel (2000) and Malpezzi & Mayo (1997) present data and comparisons, but these are generally based on simple median house prices from selected cities; more careful analysis, including estimating inter-country hedonic models, remains to be done.16 An issue of fundamental importance to future hedonic work is the collection of more and better data for hedonic estimation (as well as other kinds of housing analyses). For example, good benchmark data in many countries are still needed. Guides to improved housing data collection include Malpezzi & Mayo (1994) and Malpezzi (2000).

Hedonic Pricing Models: a Selective and Applied Review

87

Lastly, while housing is the bulk of every country’s real estate, and in fact typically over half a country’s tangible capital, hedonic models have rarely been applied to other forms of real estate. Hedonic applications to commercial real estate will be of interest for their own sake; and the functional interdependence of residential and non-residential real estate is often underappreciated by those of us focused on housing. Over the past three decades, hedonic estimation has clearly matured from a new technology to become the standard way economists deal with housing heterogeneity. Duncan Maclennan’s own work in this area, and the work of his colleagues and students, has helped push out these frontiers. Building on this progress, many exciting applications and innovations in hedonic technique undoubtedly lie ahead. I am grateful to the editors and to fellow contributors to this volume for constructive comments on a previous version. Opinions in this chapter are those of the author, and do not reﬂect the views of any other individual, or institution.

Notes (1)

(2)

(3)

(4)

In drafting this highly selective review, I have beneﬁted from many previous studies, only some of which are listed below. Even the list of other surveys is incomplete. In my own early work, I greatly beneﬁted from Ball’s (1973) classic review of early literature. Among many recent reviews, see especially Follain & Jimenez (1985a) and Sheppard (1999). Some hedonic studies use aggregate data, e.g. average levels of variables over, say, census tracts; but these have gone out of favour, partly because of aggregation bias (as discussed in Ball 1973) and partly because household/ property level data have become more readily available. The classic reference on repeat sales is Bailey et al. (1963); while there were early applications such as Nourse (1963), they were greatly popularised by several papers by Case & Shiller (e.g. 1987, 1989). Wang & Zorn (1997) provide a thorough review. There are also hybrid models that combine information from both hedonic and repeat sales estimates; see, for example, Case & Quigley (1991) and Quigley (1995). Green & Malpezzi (2001) presents further discussion of these as well as of simpler ‘models’ such as simple medians of transactions, and Laspeyres, Paasche and Divisia time series indices, and the very important user cost model of price determination. Actually, as we will see later in this section, with large samples regression techniques are used, but it amounts to the same thing.

88 (5)

(6)

(7)

(8)

(9) (10)

(11) (12)

(13)

(14)

Housing Economics and Public Policy For notational simplicity we suppress error terms. Careful consideration of house-speciﬁc errors and their ‘drift’ over time is a hallmark of Case & Shiller’s (1987) treatment. In addition to the references on repeat sales above, see Gatzlaff & Haurin (1997) and Gatzlaff, Green & Ling (1998). I also recommend the excellent review of repeat sales issues contained in Wang & Zorn (1997). Hybrid indices combine elements of two or more methods into one index. Such methods seek to take advantage of the strengths while minimising the weakness of the constituent indices. These could be time series, cross-section, or both. I have already alluded to hybrid models that combine hedonic and repeat sales methods. The essence of most hybrid models is to ‘stack’ repeat sales and hedonic models, and then to estimate the two models imposing a constraint that estimated price changes over time are equal in both models. Such methods have the advantage of making use of all available information (see Case & Quigley 1991; Quigley 1995; or Hill et al. 1999, for good examples of hybrid indices). Knight et al. (1995) use seemingly unrelated regressions as a way to get more efﬁcient coefﬁcient estimates than the coefﬁcient estimates obtained by OLS, but this procedure requires tedious matching of similar observations across years. Interestingly, a casual perusal through stacks of papers suggests that UK scholars more often cite Lancaster as their fundamental reference, and US authors more readily cite Rosen. Of course, both papers are often cited by writers from any country. See also Lancaster’s later elaboration of his ideas in Lancaster (1971). My stepsons view the problem somewhat differently, of course; in their childhood they certainly saw their rooms as refuges from noisy parents. Abraham and Hendershott and Follain use repeat sales indices, while Malpezzi combines hedonic and repeat sales indices. But whether hedonic or repeat sales indices are adopted, the general approach is the same. Without loss of generality, we’ve written one of each, when there will usually be several; or if you like, consider each (S, N, L, and C) as a vector. In fact, if housing units could, via some Star Trek-inspired machinery, be instantly and costlessly re-formed, then such costless repackaging would imply a linear structure of prices, where the dollar or pound price of each characteristic was simply added up, much as one would add different items in a shopping basket full of groceries at the checkout to obtain their ﬁnal price (Fisher & Shell 1971; Triplett 1974). It is ultimately the cost of adjustment that gives rise to the nonlinearities that are observed empirically in housing prices. As Halvorsen and Pollakowski point out, additional ﬂexibility could be built in by allowing λm to vary with each independent variable; for computational convenience and degrees of freedom, all the hedonic applications Halvorsen and Pollakowski cite, and all those I am familiar with, constrain λ to be the same across all independent variables. I have yet to ﬁnd a hedonic study of Antarctica, but that’s about the only region as yet unstudied.

Hedonic Pricing Models: a Selective and Applied Review (15)

(16)

89

Initially the indicators project was a joint World Bank–UNCHS research programme, but the World Bank has effectively ceased research in housing and urban development in recent years. See the symposium in Netherlands Journal of Housing Research (e.g. Angel, Mayo & Stephens 1993, Maclennan & Gibb, 1993, and Priemus 1992) for critical reviews of the project. See also Flood (1997). General discussion of cross-country comparisons include Malpezzi (1990), Annez and Wheaton (1984), Malpezzi and Mayo (1987), Maclennan and Gibb (1993), Harsman and Quigley (1991), Strassman (1991), and Boelhouwer and van der Heijden (1993).

6 Housing, Random Walks, Complexity and the Macroeconomy Geoffrey Meen ‘The dominant theoretical framework for microeconomic analysis of housing markets is the standard, contemporary Walrasian synthesis of the neoclassical framework. In that synthesis the behaviour of consumers and producers, and the institutions of the market itself, are modelled in an extremely reductionist framework of assumptions. Consumers and producers are assumed to be fully informed and fully rational and the system is usually regarded as competitive. Goods are assumed to be relatively simple and, in the main, either transportable to the market or capable of full, believed descriptions within the marketplace. The market itself, in which fully informed trading takes place, is essentially a single point in space which adjusts or clears in the market trading period. These reductionist assumptions, which can be relaxed, are unsurprising because the essential purpose of the theory is to show how well-functioning markets clear and how such markets interact to generate a general competitive equilibrium in an economy. ‘Although the insights of the standard neoclassical model are very important, it is clear that economic phenomena such as mis-information, commodity variety, space, time, and the nature of the market itself are all victims of the abstraction process.’ (Maclennan & Tu 1996).

Introduction The quotation above illustrates two of the concerns that run through Duncan Maclennan’s writings: ﬁrst, a dissatisfaction with conventional reductionist models that dominate mainstream housing micro and macroeconomic analysis, and second, a concern that housing markets are not aggregate, national entities, but consist of a series of interconnected local sub-markets,

Housing, Random Walks, Complexity and the Macroeconomy

91

each of which may behave in a different way. Nevertheless, Maclennan’s work recognises that housing markets have important macroeconomic consequences (see Maclennan et al. 1997, 1998b). It is the relationship between housing and the macroeconomy that is the subject of this chapter and the extent to which traditional forms of analysis have proved to be the most appropriate vehicle. In the late eighties, when a boom in UK house prices coincided with a boom in the macroeconomy as a whole, the relationship between housing and the macroeconomy became an important research topic. However, although our understanding of the interactions has improved considerably since then, it is still arguable whether this knowledge has helped us in macro policy terms since house prices themselves are difﬁcult to predict. Forecasting of house price movements, for example, has been just as poor as ever with a failure to capture important turning points, particularly during the nineties. Therefore, a major theme of this chapter is to assess where we stand at the moment and to indicate the limitations of housing research at the aggregate level. Indeed, is accurate prediction ever possible from housing models? In line with the quotation, do we need to move on from reductionist models; do we need to recognise more fully the aggregation problems implicit in macro models if local markets are the more appropriate spatial scale of analysis; is prediction impossible because housing has increasingly become like any other ﬁnancial asset, whose price follows a random walk? Do bubbles exist and imply that the coefﬁcients of housing models are unstable? More speculatively, from a totally different perspective, do ideas from complexity analysis currently in vogue in economics imply that the future of housing markets is unpredictable? If so, then knowledge that housing affects the macroeconomy is of limited value to policy makers because they cannot forecast the state of the housing market at any stage over the cycle. The chapter is divided into two substantive sections. The ﬁrst reviews the state of play on the relationship between housing and the macroeconomy and identiﬁes the key variables. Unsurprisingly, house prices play a central role, but other housing market variables such as construction have a smaller impact. The second section is devoted to what is now known about the determinants of housing markets at the macroeconomic level and, particularly, their predictability. The question of the adequacy of the standard neoclassical paradigm arises in this context and the need to take into account spatial structure. We suggest that standard, aggregate analysis, based on representative agent models, has to be treated with considerable caution in today’s climate of increasing heterogeneity and, at the very least, macro analysis has to be supplemented with careful micro analysis.

92

Housing Economics and Public Policy

The relationship between housing and the macroeconomy The literature in recent years has identiﬁed a number of potential ways in which housing markets affect the macroeconomy. These affect both the demand and supply sides of the economy. Among the most important, which are discussed in turn, are:1

•

the relationship between house prices and consumers’ expenditure;

•

the effect of housing on wages, migration, and labour markets in general. The transmission mechanisms operate in a number of ways, although house prices (both at national and regional levels) and the owner occupation rate have been suggested as particularly critical. Through labour markets, housing affects the regional dispersion of economic activity as well as the national aggregate;

•

the inﬂuence of international differences in European housing markets on monetary union;

•

the contribution of housing activity to economic cycles; and

•

the links between housing and the spatial concentration of business startups.

Housing and consumers’ expenditure In the UK, interest in the relationship between housing and the macroeconomy began in the second half of the eighties with the observation of a correlation between changes in house prices and a boom in consumers’ expenditure. Between 1986 and 1988, consumption grew in excess of 5% in each year – well above the economy’s productive capacity – whereas real house price growth averaged 13% per annum over the period. In policy terms, these movements are highly inﬂationary. However, although few researchers have doubted that a correlation exists between the two variables and the relationship appears to have continued through the nineties, considerable controversy still surrounds causality. Muellbauer (1990) was the ﬁrst to argue that a causal relationship exists between house price growth and excessive growth in consumers’ expenditure in the liberalised mortgage market conditions of that time. Some other countries, notably Scandinavia and the Netherlands, appear to have found

Housing, Random Walks, Complexity and the Macroeconomy

93

similar evidence, but there has always been a contradiction between macroeconomic time-series evidence and microeconomic ﬁndings. All forecasting institutions, including the Treasury, failed to predict the eighties consumption boom, leading to an over-relaxation of monetary policy at that time (see Meen 1996). House price booms had, of course, taken place before in the early and late seventies, but no corresponding boom had occurred in consumption. However, the difference between the periods was the liberalisation of mortgage markets in the early eighties. Under the rationed conditions of the seventies, an increase in house prices could not be translated into higher consumption through borrowing. But the removal of rationing enabled households to borrow against the accumulated equity in their houses to ﬁnance consumption. Furthermore, liberalisation of credit markets increases the sensitivity of housing markets and, indirectly, consumption to changes in interest rates. Equity withdrawal – the propensity to borrow on mortgage more than is required to ﬁnance the purchase of a home – rose dramatically in the late eighties, reaching almost 7% of household income in 1988. Similarly, Boelhouwer (2000) reports that, in the Netherlands, rapidly increasing house prices have been responsible for sharply increasing consumers’ expenditure in 1999 through the same route of equity withdrawal. But, although equity withdrawal began to re-emerge in the UK in the late nineties, the volume has not approached that in the late eighties. Furthermore, equity withdrawal disappeared altogether in the early nineties as house prices collapsed. Negative equity replaced equity withdrawal and Henley (1998) has argued, reasonably, that negative equity lowered labour mobility in the nineties. However, the boom in consumption of the second half of the eighties has also been attributed to other causes. King (1990) argues that, instead, the boom represented a re-assessment of permanent income. Since the demand for housing, and hence house prices, would also be affected by the revision to expectations, the correlation between consumption and house prices may represent a proxy effect. In terms of policy, which explanation (if either) is true is important. The explanation in terms of a wealth effect might suggest tax measures explicitly targeted at the housing market to prevent price booms – a measure that would be irrelevant if expectations were the true causal mechanism. As Hort (1997) points out, macroeconomic and microeconomic data yield conﬂicting evidence. On microeconomic data, both Attanasio & Weber (1994) and Miles (1997) do not support explanations of the consumption

94

Housing Economics and Public Policy

boom in terms of house price changes. But on macroeconomic data, Muellbauer & Murphy (1990) ﬁnd signiﬁcant effects in the UK. Berg (1994), Berg & Bergstrom (1995) and Koskela & Viren (1992) ﬁnd that in the Scandinavian countries, which have experienced house price movements similar to the UK, there are signiﬁcant effects. In the USA, Case et al. (2001) have found signiﬁcant housing wealth effects in a panel study across the US states and over fourteen countries. Overall, however, the inconsistency between micro and macro results remains a puzzle.

Housing and labour markets In principle, the effects of housing on labour markets are straightforward. However, compared with the last sub-section, which was concerned with aggregate demand, any effects on the labour market affect primarily aggregate supply. Therefore the inﬂuences on the economy are more likely to be permanent than transitory. First, since housing costs are a component of consumer prices, a direct effect on wage pressures is likely and is relatively uncontroversial. Bover et al. (1989), Blackaby & Manning (1992), and more recently Cameron & Muellbauer (2000) all ﬁnd house prices to be important in determining wage claims. Furthermore, the last of these studies shows that prices have a permanent effect on the level of unemployment. At a more local level, housing constraints and high housing costs are the central element of the key worker shortages facing South East England (see Monk 2000a). Second, regional variations in house prices may have an impact on migration patterns, and thus on regional unemployment rates. As such they contribute to the North/South divide. It has been argued that high house prices in southern England act as a barrier to mobility as workers seek to move from high unemployment northern areas to low unemployment southern regions. This effect is likely to have been particularly important in the late eighties when the ratio of house prices in the South East relative to the North reached record levels (see Meen 2001, chapter 2). At the time of writing, recent increases in house prices, concentrated in the South, have again exacerbated the divide. There is, however, some doubt as to the quantitative impact of relative house prices on migration patterns. The earliest work on migration in the UK (a series of papers by Hughes & McCormick 1981, 1985, 1987, 1990, 2000) was concerned primarily with the effect of tenure on migration rather than house prices. But in terms of house prices, Hughes & McCormick indicate that the

Housing, Random Walks, Complexity and the Macroeconomy

95

effect of relative house prices on out-migration rates is negligible and the inﬂuence on destination choice is only small (Hughes & McCormick 1990, p. 103), although they note that the estimates are obtained for members of the labour force and the retired may have a higher propensity to migrate from the South East to take advantage of accrued capital gains. But Hughes & McCormick’s ﬁndings are by no means universally accepted and Thomas (1993), for example, suggests that, after standardisation for the moving motivation is made, relative house prices have a strong effect on destination choice. On balance, however, as Gordon (1990) argues, it is unlikely that housing price variables alone can account for low rates of migration in the UK, particularly among manual workers (who have lower rates of owner occupation). Cameron & Muellbauer (2000) also ﬁnd that relative house prices primarily affect the wages of non-manual workers.

Housing and monetary union Maclennan has been among those who have argued strongly that differences in housing conditions and institutions across Europe raise problems for monetary union (see Maclennan et al. 1998b). Considerable differences in housing volatility still exist across Europe, both in terms of house prices and construction activity (see Stephens 2000 and Ball & Wood 1999), despite the presence of a common monetary policy since the beginning of 1999 for most European Union members. At the time of writing, only the UK, Sweden and Denmark have opted out. The most important differences in housing structures across Europe concern: • variations in owner occupation rates; • differences in mortgage debt as a percentage of GDP; and • the fact that in some countries mortgage debt is primarily at ﬁxed rates but in others debt is mainly variable rate.2 In the UK it is sometimes argued that owner-occupier rates are high and this is certainly the case compared with Germany and France. The owner occupancy rate in Britain stands at 69%, but only approximately 40% and 55% in Germany and France respectively. But this is relatively modest compared with Ireland and Spain at around 80% and Greece and Italy at approximately 75%. However, high owner occupancy rates are not necessarily correlated with high levels of outstanding debt. The ratio of outstanding mortgage loans

96

Housing Economics and Public Policy

to GDP stood at 57% in 1995 in the UK, only exceeded in Europe by the Netherlands at 60% and Denmark at 65%. But of the high owner occupancy countries, the debt ratio was only 27% in Ireland, 22% in Spain, and 6% in Greece and in Italy (Stephens 2000). Clearly a high degree of development in ﬁnancial markets is not a necessary condition for high levels of owner occupation. Therefore, in low debt countries, variations in interest rates are not critical despite high owner occupancy rates. But, even at high debt ratios, variations in market interest rates would not be signiﬁcant for the macroeconomy if the majority of mortgage debt were at ﬁxed rates. Around 70% of mortgage debt is variable rate in the UK, but only 10% in the Netherlands, 20% in France and 40% in Germany. Although it is certainly not the case that housing conditions are uniform across the rest of Europe, the United Kingdom has experienced a special combination of circumstances – high owner occupation rates, high debt levels and high proportions of variable rate debt. In these circumstances, it would be unsurprising to ﬁnd that (i) activity in mainland Europe is less elastic in response to changes in interest rates3 and (ii) the same level of interest rates may not be suitable to all European countries.

Housing construction, growth and economic cycles The previous sections emphasised the interactions between house prices and the macroeconomy. Far less research has been conducted on the effects of new housing construction, although housing pressure groups have long argued that one of the reasons for increasing housing investment is the positive beneﬁts to GDP and employment (see, for example, Clapham 1996). But housing construction, potentially, plays a number of roles in the national economy. In addition to adding to aggregate demand, perhaps, the two key contributions are:

•

as a factor of production; and

•

as a contributor to cycles in the economy.

As a factor of production, in principle, housing has three effects: (1) directly by increasing the capital stock; (2) by affecting the rate of technical progress: for example, well-housed workers and children are more productive and educational standards rise; and

Housing, Random Walks, Complexity and the Macroeconomy

97

(3) by either crowding-in or crowding-out other parts of the capital stock, i.e. extra housing may reduce business investment and, therefore, output growth in the economy. Nearly all empirical work has concerned the effects on demand and crowding-out. But since new housing construction is only a small part of the economy, its direct impact is limited. New construction annually comprises only approximately 1% of the housing stock and private and public housing investment totalled only 2% of Total Final Expenditure in 1999. As an example, Meen et al. (2001) use an econometric model to simulate the economic effects of reducing new housing construction by 25% per annum over a 10-year period. Although precise numbers from econometric models are always subject to margins of error, certain points stood out.

•

Though cutting housing production would reduce demand and GDP in the ﬁrst few years, through standard multiplier effects, even the initial effects are modest.

•

Crowding-out causes any effects to die away over time. Although employment in the construction industry falls substantially, in the long run, there is little effect on unemployment in the economy as a whole. Employment in other industries expands.

But these conclusions may be misleading because they treat housing in the same manner as any other consumption good, ignoring the importance of housing as a capital good. In other words, they omit points (1) and (2) above. In general, standard analysis of housing and the macroeconomy ignores any induced enhancements to productivity or technical progress. In the current stage of knowledge, we simply do not know whether they are signiﬁcant or not. Even if additional housing construction has limited long-run effects on the national economy, housing is one of the most volatile sectors of the economy and, clearly, volatility causes problems for the design of macroeconomic policy. Neither is the volatility of new housing conﬁned to the UK. Ball & Wood (1999), for example, discuss both long-run trends and volatility over a wide range of countries, covering data from the nineteenth century onwards. They suggest that, since the 1970s, ﬂuctuations in housing investment have added to instability in the world economy. Meen (2000) explores the reasons for this volatility in UK housing markets and suggests that monetary policy, designed with general inﬂation targets

98

Housing Economics and Public Policy

in mind, in some circumstances is inconsistent with stability in housing, if housing is inelastic in supply. Policy exacerbates ﬂuctuations in housing.

Housing and business start-ups Although the previous sub-section has suggested that housing may crowd out business investment, an alternative line of thought has suggested that, in fact, the two are positively related. First, Black et al. (1996) suggest that house values are important in providing collateral for the start-up of new businesses, although the view is challenged by Cressy (1996). Second, although the determinants of industrial location are multi-faceted, surveys suggest that the availability of skilled labour is one of the more important factors determining industrial ﬁrm location. But skilled labour is attracted to areas of better housing. Therefore, both ﬁrms and the skilled workforce tend to be co-located. It appears that innovating industry is attracted to highly qualiﬁed labour, which, in turn, is attracted by high-quality living conditions, typically in the home counties of South East England. Furthermore, Keeble & Walker (1994) show that the key variable determining new ﬁrm formation is the rate of population growth. They argue that population growth incorporates both demand inﬂuences (through the provision of local markets) and supply effects. Increases in population raise the pool of potential entrepreneurs, which, they argue, is stimulated by residential preferences. Furthermore, South East England has a high rate of new ﬁrm formation compared with other regions, particularly in the innovative, high growth industries. This performance appears to be related to the disproportionate presence of highly skilled technical, managerial and professional groups in the South East from whom potential entrepreneurs principally emerge. Since these entrepreneurs start new ﬁrms close to where they live, areas of strong economic growth, high-quality environments and amenities tend to be self-reinforcing (see Barkham 1992).

Are house prices predictable? The previous section illustrates that, in terms of the relationship between housing and the macroeconomy, prices are the most important inﬂuence. New construction has a more limited (although possibly understated) effect. Although other aggregate housing indicators are regularly reported – notably housing transactions, arrears and possessions – their inﬂuence on the macro-

Housing, Random Walks, Complexity and the Macroeconomy

99

economy is rarely discussed. A possible exception is mortgage advances, but again the main effect is generally taken to be through house prices. It is, however, worth noting that models of house prices have had a chequered career in the UK in terms of prediction. The key issue in this section is to examine whether standard models based on the neoclassical paradigm have proved adequate or whether we have to go down other routes as might be suggested by the opening quotation. In order to understand the problems, we need to examine both the methods that are currently typically used and to ask whether house prices are intrinsically unpredictable. Clearly, if house prices are unpredictable, then knowledge of the relationship between prices and the macroeconomy is of limited value to policy.

House price to income ratios By far the most common method of analysis used in the UK to predict future house price movements is a simple rule-of-thumb related to the ratio of house prices to household income. The rule-of-thumb states that if the current ratio is above the long-run average (usually assumed to be constant), future prices will fall and vice versa. No theory lies behind the ‘analysis’, which can be highly misleading. Figure 6.1, at ﬁrst sight, seems to suggest that a constant ratio has existed since the early seventies. The ratio rises in boom periods and falls in the slumps. Two recent events, however, illustrate the difﬁculties. First, the UK economy began to recover from a slump, with GDP increasing from 1993; based on the ratio, commentators predicted that house prices would also begin to recover. In fact, the ratio fell further with prices not rising until the end of 1996. The fall in the ratio over this period can be seen in Fig. 6.1. Second, although the period 1998–2000 has been considered a boom period in terms of house prices, the ratio of prices to incomes has risen modestly compared with some previous booms. The question is why prices are not even higher than we currently observe, particularly given the very low levels of mortgage interest rates. Therefore, although there may appear to be some simple long-run ratio, if we look in more detail, the ratio has been far from constant in the long run. Those who made predictions on the basis of this ratio have made serious errors, particularly since 1990. A number of issues that arise in house price modelling and prediction undermine simple house price to income ratio analysis. Although such ratios have the advantage of simplicity and are readily explainable to a wider audience,4 they are seriously misleading.

100

Housing Economics and Public Policy

180 1995=100 160

140

120

100

80 74 76 78 80 82 84 86 88 90 92 94 96 98 00 Fig. 6.1 UK ratio of house prices to personal disposable income.

A theoretical framework Although all theoretical and empirical research on housing demand and house prices stresses the importance of income, few studies imply that income is the sole determinant. However, there is an increasing consensus as to the most appropriate modelling approach to house prices. In line with the opening quotation, the consensus is almost entirely neoclassical. Although there have certainly been cries that this is an inappropriate framework, in practice it is dominant. Empirical models of house prices are generally derived as representative agent models, where households maximise utility subject to a lifetime budget constraint. From this approach, equation (6.1) can be derived as the basic house price relationship underlying most modern work. (6.1) where: g(t)

= real purchase price of dwellings

R(t)

= real imputed rental = household marginal tax rate

i(t)

= market interest rate

Housing, Random Walks, Complexity and the Macroeconomy

101

= depreciation rate on housing = general inﬂation rate (.)

= time derivative

g·e / g(t)

= expected real capital gain.

Equation (6.1) implies that the return on housing, represented by the imputed rental plus any capital gain on housing net of depreciation, equals the post-tax return on alternative assets. In principle, data permitting, we could simply use equation (6.1) as our predictor of house price movements. But note: (1) income has no role in this speciﬁcation; (2) it treats housing in exactly the same way as any ﬁnancial asset; (3) there are no lags in the speciﬁcation; the market is fully efﬁcient and prices react immediately to, say, changes in interest rates; and (4) in practice, with the exception of Meese & Wallace (1994), no empirical model estimates equation (6.1) directly. Equation (6.1) could be seen as the most extreme version of the neoclassical model. In practice, empirical models in the UK (and in the US) pay only lipservice to equation (6.1). This is partly due to data deﬁciencies – we cannot measure R(t) directly – and partly due to rather sloppy analysis. In the UK, a more typical example looks like equation (6.2), into which the expected determinants of R(t) are substituted. Notice that income now reappears in the speciﬁcation, but (6.2) only bears a loose relationship to the underlying theory. Almost all empirical speciﬁcations can be justiﬁed by the inclusion of different variables in (6.3).5 (6.2) (6.3) where: RY

= real personal disposable income

W

= real wealth

102

Housing Economics and Public Policy

HH

= number of households

H

= housing stock

In (6.2) a number of factors, in addition to income, affect prices – real interest rates, taxation, demographics, the supply of housing, wealth and expectations. If any of these variables change, house price to income ratios will be inadequate as a predictor.

Efﬁciency and random walks in housing Life-cycle models under rational expectations imply that house prices follow a random walk. In this framework, housing behaves in the same way as a ﬁnancial asset that denies many of the distinguishing characteristics of housing. As such, prices are efﬁcient and unpredictable, because lagged values of no variable have an inﬂuence on current house price movements. Clearly, if this is a true representation of price movements, this causes considerable problems for macroeconomic policy. However almost all evidence suggests that housing markets are neither efﬁcient nor can be adequately characterised as a random walk (see, for example, the surveys of US evidence by Gatzlaff & Tirtiroglu 1995 and Cho 1996). Internationally, the most comprehensive set of efﬁciency tests have been conducted by Englund and Ioannides (1997). Following the approach for the USA of Case & Shiller (1989, 1990), the basic equation is given by (6.4): (6.4) where: gt

= real house prices at time (t)

εt

= error term

∆

= ﬁrst difference operator

ln

= natural logarithm of the variable

If housing markets are fully efﬁcient, then no autocorrelation in price inﬂation is expected, (βj) = 0, for all (j) so that house prices follow a random walk. Englund and Ioannides, estimating a panel model across 15 countries, ﬁnd that real house price inﬂation exhibits short-run positive autocorrelation

Housing, Random Walks, Complexity and the Macroeconomy

103

and longer-term mean reversion. Furthermore house prices are predictable by lagged income growth and real interest rates. An alternative approach is to test for the presence of signiﬁcant lagged variables in equations such as (6.2). In an efﬁcient market, prices should respond immediately to changes in the determinants so that lags should have no effect. In European studies, tests have generally been conducted in the context of error correction models. Without exception, lags are found to be important and, indeed, prices appear to adjust fairly slowly to equilibrium following any shock. Using rather different methods, similar results are found for the USA. DiPasquale & Wheaton (1994) and Mankiw & Weil (1989) ﬁnd support for the view that housing markets are not fully efﬁcient. If housing markets are not efﬁcient, then what are the key differences from ﬁnancial markets that can account for the presence of lagged adjustment? There appear to be two main factors: ﬁrst, the presence of transactions costs which are much larger than, say, for shares (see the chapter in this volume by Quigley) and, second, the presence of credit market constraints. Whereas it may be reasonable to assume that large institutions operating in ﬁnancial markets face perfect credit markets, young ﬁrst-time buyers entering the owner-occupier market do not. Because of adverse selection, for example, they face downpayment constraints (see Haurin et al. 1997). This has two implications. First, prices become more volatile and regional diversity becomes more pronounced. Second, the presence of ‘hurdles’ means that households remain for long periods of time ‘off their demand curves’. When the hurdles are overcome, housing adjustment is more lumpy than in the presence of perfect capital markets. Again this adds to volatility in the housing market. In summary, there are good reasons why simply applying the standard neoclassical life-cycle model to housing is inadequate. At the very least, transactions costs and credit markets have to become a fundamental part of the analysis. Lagged adjustment is to be expected in an extended framework implying that, in principle, house price movements should be partly predictable. We cannot appeal to random walks as an explanation of our failure to predict adequately future house price movements. However, there are a number of other reasons that impinge upon our ability to predict price changes. One of the most important is the fact that the neoclassical model is based on a single representative agent. No consideration is given to the aggregation conditions involved in moving to the macro level. As I show in the next section, this issue has become very important in recent years, where agents have become increasingly heterogeneous.

104

Housing Economics and Public Policy

Aggregation and structural change in the housing market Models based on equation (6.2), incorporating lags, have performed poorly during the nineties. As an example, Fig. 6.2 shows the post-sample errors of the house price equation estimated in Meen (1990). The record is hardly distinguished with consistent over-prediction of the outcome for price changes after 1990. It has been suggested that housing markets are so volatile and difﬁcult to predict that variable parameter methods are necessary to model house prices. Methods of estimation have been imported from the ﬁnancial econometrics literature. But volatility in itself does not necessarily imply non-constant parameters. Volatility may simply arise from high elasticities with constant parameters. In fact, the empirical evidence (Meen 2001 Chapter 6) suggests that house prices are particularly sensitive to changes in interest rates and incomes.6 But variable parameter methods may be necessary if the coefﬁcients are not constant over time (see Hall et al. 1997; Brown et al. 1997). This might be the case if bubbles occur regularly.7 In fact, I would argue that the parameters of house price equations are no more unstable than many macroeconomic time-series; supposed instability is sometimes used as an excuse for inadequate theoretical speciﬁcation and detailed analysis of the micro structure of markets. In the UK there appear to have been only two major structural changes corresponding to explicit policy initiatives. The ﬁrst took place in the mideighties in response to ﬁnancial deregulation. For aggregate housing market

0.08 % 0.04

0.00

-0.04

-0.08

-0.12 70 72 74 76 78 80 82 84 86 88 90 92 94 96 Fig. 6.2 House price prediction errors. Source: Meen and Andrew (1998).

Housing, Random Walks, Complexity and the Macroeconomy

105

analysis, its main impact was that house prices became more sensitive to changes in interest rates (see Meen 1996). However, if allowance is made for this change, there is little evidence that the remaining coefﬁcients of house price equations altered signiﬁcantly. The second, in the early nineties, has proved more difﬁcult to model. The extent of the problem was shown by the errors in Fig. 6.2. The two most likely explanations of the breakdown in the relationship concern (i) changes to housing market risk in a new low inﬂation environment and (ii) the effect of structural labour market changes. Meen & Andrew (1998) argue that the latter was probably the more important, leading to a form of aggregation bias discussed in more detail below. It highlights the need to supplement macroeconomic analysis with detailed analysis of underlying microeconomic movements. However, it also suggests that variable parameter modelling is a rather poor substitute for looking at the fundamental causes of change, particularly if the interest is in policy implications. Neoclassical models are based on the analysis of representative agents. Aggregate models are then constructed by simply summing individual functions. But the conditions under which this is a valid procedure are strict; either all agents must be identical in their behaviour (homogeneous) – characterised by all the coefﬁcients of the individual agent functions being identical8 – or movements in the regressors in the functions must be the same over time. For example, everyone’s income must grow at the same rate over time. More recent research by Hall et al. (1999) modiﬁes the condition to one in which the regressors follow common stochastic trends. Similar conditions exist in terms of spatial aggregation and reﬂect Maclennan’s concern with the importance of local housing markets. For modelling at the national level to be valid, all sub-national markets must behave in the same way or economic conditions within those spatial markets must be identical. Clearly these conditions are rarely likely to hold. It is easy to think of cases in which the individual agent aggregation conditions will not hold. First, the parameters of housing demand functions (and hence price equations) are likely to vary by demographic group. For example, the income and price elasticities of housing demand by young, ﬁrst-time buyers are unlikely to be identical to those who are retired. Second, Holmans (1996) shows that, during the nineties, there was a strong shift in demand by the 20–25 age group away from owner occupation towards the private rented sector. Meen (2001) suggests that this was associated with a fall in relative real income growth by young households compared with older age groups, i.e. a violation of the second aggregation condition through a change in the income distribution.

106

Housing Economics and Public Policy

A priori we might expect that these changes would lead to an over-prediction of house price growth in the nineties by ‘traditional’ models, consistent with Fig. 6.2. Similarly, the failure of national models in the nineties reﬂects differences in behaviour at the regional level. Meen (1999b) shows that there are nonrandom differences in the coefﬁcients of regional house price equations; typically the south of the country is more sensitive to national economic stimuli (particularly interest rate changes) than the north of the country. I speculate in that paper that these differences reﬂect variations in debt gearing around the country.9 The recent continuing low levels of interest rates, therefore, confer particular beneﬁts to the South, contributing to the increasing regional dispersion in house prices. Furthermore economic growth has varied considerably across the regions in the nineties; although the slump hit the South hard compared with earlier recessions, growth since the midnineties has also been concentrated on the South. In conclusion, in the modern environment, characterised by increasing dispersion, prediction and policy analysis at the macro level have to be backed up with careful microanalysis. Heterogeneity matters and is, typically, not adequately captured in simple representative agent models.

Complexity and housing markets Ideas developed from complexity are becoming increasingly important in economics but, although regional scientists have incorporated the techniques into their models of city structures, the impact of complexity on empirical housing models has, so far, been small. However, one of the implications of complex systems is that events are typically unpredictable, although for very different reasons from those employed in random walk models. Complex systems are those with a very large number of interacting parts where the interactions are non-linear and the behaviour of the system as a whole cannot be understood simply from a consideration of the constituent elements. A central feature of complexity is self-organisation. Seemingly individual action generates beautiful, complex and ordered patterns of behaviour at an aggregate level. However, modelling housing behaviour using complexity is fundamentally different from traditional neoclassical econometric models; in conventional models, all agents are homogeneous and random errors are considered a nuisance. By contrast, the properties of modern agent-based computation

Housing, Random Walks, Complexity and the Macroeconomy

107

models, often used to apply economic principles to complexity, are dependent on stochastic distributions, stressing the heterogeneity between agents. The driving forces and structures, therefore, are very different from the neoclassical paradigm. Multiple equilibria are the norm; furthermore, although equilibria may exist, they may not be achievable or may be unstable. Lastly, although equilibria may be obtainable asymptotically, the system may remain out of equilibrium for very long periods of time. The dynamics are typically more interesting than any long-run equilibrium. This is a highly attractive approach, particularly for modelling local housing markets where interactions between agents are strong, but such models have important aggregate properties. From our point of view, a key feature of these models is that they exhibit ‘tipping’ or thresholds that makes prediction in these markets extremely difﬁcult. Complexity suggests that housing markets may organise themselves into a state ‘on the edge of chaos’. In this case small random shocks to individual agents may generate completely different patterns of spatial (and aggregate) housing demand. The system may appear to be in equilibrium for many years (and over this range is predictable), but what we are, in fact, observing is a drawn-out disequilibrium state, which suddenly tips very quickly to a new state through a series of random events that are unpredictable. Using a totally different framework from the random walk model, which assumes complete rationality and individual homogeneity, prices become unpredictable in a world of bounded rationality and heterogeneous agents. Models of this form are not just an intellectual curiosity; work by Galster et al. (2000) suggests that thresholds are an important feature of local housing markets and might be used to explain why, for example, local hot-spots emerge, which can co-exist with poorer areas (see also the chapter by Galster in this volume). Forecasts for local housing markets are becoming increasingly common in this country, but complexity suggests that the foundations on which they are based can be insecure. However the implications for aggregate housing market forecasts are not yet entirely clear. Disequilibrium states can persist for long periods and, given path dependency and lock-in, there is some basis for prediction. However, we should not be surprised when highly non-linear behaviour occurs and traditional aggregate, neoclassical-based relationships break down.

Conclusions Analysis of the interactions between housing and the macroeconomy has come a long way over the last 15 years. Theoretical and empirical evidence

108

Housing Economics and Public Policy

has been provided for both the demand and supply sides of the economy; the former has concentrated primarily on consumers’ expenditure, whereas the latter has centred on wage formation, migration and unemployment. Work has been conducted at both the national and regional levels. In this work, house prices play the central role, although the impact of new construction in the economy is a neglected ﬁeld. Therefore, in macroeconomic policy terms, an understanding of the determinants of price movements is central. In this chapter, I have suggested that house price analysis is currently conducted in two frameworks. First by using rules-of-thumb based on simple house price to income ratios; such analysis is highly misleading and should be treated with a great deal of caution. Second, most modern academic work is based on neoclassical models set in a life-cycle framework. Although I have been critical of some aspects of this paradigm, one should not lose sight of the valuable insights into the working of the housing market that the approach has generated. It is likely to remain the dominant paradigm. However, it is right to be concerned with the uncritical use of the neoclassical model. In particular, I have highlighted two issues that are important in the current environment: (i) the use of representative agent models in a world of increasing dispersion both over household types and space; this leads to problems of aggregation; and (ii) the fact that there are alternative approaches, which yield insights that the neoclassical model cannot; models of complex systems lead to highly non-linear behaviour, multiple equilibria and outcomes that are typiﬁed by long-lasting disequilibria. Although, in the current state of knowledge, one cannot conclude that complexity approaches are always more appropriate than neoclassical analysis, there are instances when they are more likely to be appropriate. Finally, the warnings of Maclennan and Tu in the opening quotation are valid. It is clear that the reductionist approach of the standard neoclassical model is not always the most appropriate to housing analysis, although in many cases it will be. But we need to be aware of the very different implications that arise at both micro and macro levels from adopting a different approach. A slavish following of one approach is rarely likely to be optimal.

Notes (1) Planning also has a key inﬂuence on both national and regional economies, but planning is discussed elsewhere in this volume (in the chapter by Bramley). (2) Planning systems also vary across Europe but these are not discussed.

Housing, Random Walks, Complexity and the Macroeconomy

109

(3) Note, however, that even this is controversial. Britton & Whitley (1997) estimate simple IS/LM models for different European economies and ﬁnd no evidence that the IS curve is more elastic with respect to interest rates in the UK. (4) House price trends attract a huge amount of media attention. The virtues of price to income ratios are clearly important in this context. (5) A high degree of variation in the regressors also appears in the quantitative US literature; see, for example, DiPasquale and Wheaton (1994), Abraham and Hendershott (1996), Malpezzi (1999). (6) Note that this is one of the reasons that forecasting is so difﬁcult. Small errors in income or interest rate forecasts generate large errors in house price forecasts. (7) There are dangers in assuming that because bubbles exist in ﬁnancial markets they automatically occur in housing markets. The parallels are not exact; for example, housing bubbles may be ameliorated by credit market constraints. Bubbles may, and at times probably do, exist but there are alternative explanations for price movements that look like bubbles. Although not discussed in detail here, credit multipliers generate similar behaviour. Furthermore, under complexity, increasing returns and ‘herd’ behaviour can generate outcomes that mimic bubbles. One needs to be wary of assuming that strong volatility in prices necessarily signals the presence of bubbles. Forming appropriate tests is far from straightforward. Tests for the existence of time-varying parameters in conventional house price equations are far from adequate. (8) Pesaran and Smith (1995) show that the parameters of aggregate functions are inconsistent except under coefﬁcient homogeneity. (9) Note that debt gearing also affects the volatility of house prices and credit multiplier models can be used to explain the strong cycles in house prices (see Meen 2001).

7 Taxation, Subsidies and Housing Markets Gavin A. Wood

Introduction One of the most important inﬂuences on my academic work in housing economics has been the idea that various frictions such as space, information asymmetries, regulation and ﬁscal measures can impede the operation of efﬁcient housing market processes. Segmentation of housing markets can be a by-product of these frictions. That this idea has been so inﬂuential is in large part due to collaboration with Duncan Maclennan over a period that is approaching 25 years. In this chapter I investigate the role of taxation as a ﬁscal intervention that impacts on housing market processes. The literature on the distributional and allocative impacts of taxation is reviewed. The role taxation plays as one possible cause of housing market segmentation is a particular focus of the chapter (see also the chapter by Yong Tu in this volume). My own recently completed empirical work on this subject is summarised, and used to highlight the possible role of clientele effects as a source of segmentation in private rental markets.1 An interesting aspect of taxation measures and housing is that such measures are rarely viewed as an explicit component of housing programmes designed to further housing policy goals2 (Rosen 1985; Bourassa & Grigsby 2000). Typically the direct and indirect tax base and schedule of tax rates are framed with broader ﬁscal policy goals in mind. Their impacts on the price of housing in different tenures, or the price paid by households with different incomes, is a relatively less important inﬂuence in this regard. Yet the budgetary impacts of ‘ofﬁcial’ housing policy can be overwhelmed by taxation measures (Yates & Flood 1987). Their signiﬁcance in housing policy debates is in large part due to the widely held view that the taxation of housing is the source of distortions and inequity. As a result of tax measures:

Taxation, Subsidies and Housing Markets

111

•

the price that must be paid in the owner-occupied housing market for a given bundle of housing services will vary inversely with the income of the purchaser;

•

the price paid by homeowners will typically fall as they age; and

•

the cost of supplying rental housing is inversely related to the income of the landlord.

In the next section I explore the distributional issues that arise as a result of these distortions. Evidence from micro data is presented to show that lowincome home purchasers with a correspondingly low net worth are particularly disadvantaged. A more novel ﬁnding is that taxation of landlords can result in the concentration of marginal (high cost) investors in low-income rental housing. Thus the low-income tenants of affordable rental housing end up paying rents that are high relative to the capital values of the property they occupy. I then analyse the impacts of tax measures on tenure choice using tax arbitrage models. It is shown that when agency costs (the costs of managing landlord–tenant relationships) are taken into account, home ownership has a relative after-tax cost advantage over renting for most housing consumers. Renting is only ﬁnancially advantageous to low-income individuals, but this is typically restricted to high value segments of the market that are not affordable; because marginal landlords are concentrated in low-income rental housing, such higher value housing is ﬁnancially unattractive to lowincome households unless targeted subsidies overcome homeownership’s relative cost advantage in this segment of the housing stock. The next section turns to questions of efﬁciency. The early literature concentrated on how homeowners beneﬁt from tax subsidies that are the source of static deadweight welfare losses. Since Laidler’s (1969) seminal contribution there have been numerous attempts to measure such losses. However, there is a parallel literature that recognises that the housing sector as a whole is tax advantaged (Vandell 2000). Thus over-investment in housing is encouraged, and economy-wide productivity losses can ensue unless offset by positive externalities.3 I review this literature and discuss other potential sources of inefﬁciency, such as tax clientele effects and lock-in effects, which are less well known but can be important sources of housing market segmentation. A concluding section offers some comments on policy implications.

112

Housing Economics and Public Policy

Distributional issues Vertical and horizontal equity in owner-occupied housing markets OECD countries have a long tradition of extending tax preferences to owneroccupied housing (Wood 1988b). Among the English-speaking OECD countries there is an increasingly uniform tax treatment. Australia, Canada, New Zealand and the United Kingdom exempt imputed rental income and capital gains from taxable income, but do not permit deduction of mortgage interest and real estate taxes from taxable income (Bourassa & Grigsby 2000). The USA grants very favourable tax treatment because it not only exempts imputed rental income, it also permits deduction of mortgage interest and real estate taxes.4 Why do these tax arrangements raise equity concerns? The main reasons are that they generate a non-neutral intra-tenure and inter-tenure tax treatment of households, with the consequence that the price paid for a given bundle of housing services is a function of the occupants’ incomes, and the tenure in which the housing services are consumed.5 I focus in this section on distributional concerns that arise because of variation in the price paid for owner-occupied housing. To assist analysis I begin by describing an intra-tenure neutral tax treatment of owner-occupied housing. It would treat income from imputed rents and capital gains in a symmetric fashion; it would permit the deduction of all operating expenses, including interest and real estate taxes, from imputed rents; and it would permit the deduction of capital improvement and transaction costs from taxable capital gains. To demonstrate formally the tax neutral properties of these provisions, consider an owner-occupier who has acquired housing capital (q).6 It is assumed that the owner-occupier’s holding period (T) is exogenous; house prices and rents appreciate at the constant uniform rate πh;7 the constant rate of general inﬂation (π) may differ. The present value of the stream of after-tax returns (V) is given by:

(7.1)

where p(0) is the asset price of housing in year zero, m is mortgage debt that is not amortised, τ is the marginal income tax rate, r(t)q is gross imputed rent, υ is operating cost (including real estate taxes) as a fraction of asset price and i is the market rate of interest at which investors can borrow or lend. The

Taxation, Subsidies and Housing Markets

113

parameter k is the owner-occupier’s discount rate, and δ = πh – k.8 In competitive markets the present value function can be set equal to zero. Evaluating the integrals, and factoring q and m we obtain two sufﬁcient conditions:

(7.2)

(7.3) Combining equations (7.2) and (7.3) we ﬁnd that

(7.4)

The right-hand side of equation (7.4) represents the after-tax annual costs of the homeowner, or his/her user cost of housing capital. It is a measure of the effective price that is paid by a homeowner, including the after-tax cost of equity capital.9 The neutrality of this simple expression is evident. The price paid by the homeowner is independent of its marginal income tax rate and the chosen loan-value ratio. Consider now the typical tax treatment of owner-occupied housing in English-speaking OECD countries.10 The present value function can now be deﬁned as: (7.5) Repeating the steps described in equations (7.2) and (7.3) we obtain: (7.6) where α is the loan-value ratio. The ﬁrst term on the right-hand side user cost expression is the homeowner’s ﬁnancing costs if 100% equity ﬁnanced, and given the tax exempt status of imputed rents. The second term is the tax penalty for a homeowner using debt ﬁnance for the fraction α of acquisition costs. Mortgage interest is not deductible under the typical tax arrangements, hence the tax penalty. However, there is an offsetting component (the third term) due to the erosion of the real value of outstanding mortgage debt, which represents an untaxed addition to the household’s resources.11

114

Housing Economics and Public Policy

It is evident from equation (7.6) that user costs are reduced by the typical tax treatment of homeowners. If we subtract equation (7.6) from equation (7.4) and express as a proportion of the neutral tax user cost measure, we obtain the effective rate of subsidy (S), which is given by:

(7.7) The non-neutral tax treatment lowers the effective price of owner-occupied housing by an amount that is directly related to the owner’s marginal income tax rate. Table 7.1 computes S for alternative values of homeowners’ marginal income tax rates and loan to value ratios. At each of the chosen loan to value ratios, S declines at the lower marginal income tax rates to the extent that it is less than one half its value at the highest marginal income tax rate. At low marginal tax rates S declines marginally as the loan to value ratio increases, but this pattern is reversed at the highest marginal income tax rate.12 Concrete evidence of the distortionary effects of taxation arrangements is presented in Table 7.2, which lists estimates of the user costs of 9908 Australian homeowner households using the 1990 Australian Bureau of Statistics Income and Housing Cost Survey. The households have been ordered into deciles according to household income, and their marginal tax rates have been estimated following the method advocated by Hendershott & Slemrod (1983).13 We ﬁnd that mean user cost in the lowest income decile is 1.8 times that typical among households in the highest income decile, a multiple which is generated by a 3.21 percentage point difference in user cost. If the typical household in the lowest income decile were to beneﬁt from a

Table 7.1 The effective rate of subsidy (%) by typical loan to value ratios and marginal income tax rates.1 Marginal income tax rates Loan to value ratio

21%

38%

47%

0.25 0.50 0.75

28 26 24

56 55 54

72 74 75

Note 1 These effective rates of subsidy are computed using an interest rate of 10%, a rate of house price appreciation of 5.75% and operating costs that are 2.4% of asset value.

Taxation, Subsidies and Housing Markets

115

Table 7.2 Mean household income, house value, marginal tax rate, loan-value ratios and user costs by decile of household income inclusive of net imputed rent for all owning households.1

Decile

Household income (A$)2 (1)

House value (A$) (2)

Marginal tax rate % (3)

Loan-value ratio % (4)

User cost % (5)

1 2 3 4 5 6 7 8 9 10 All

12 752 19 952 25 345 30 888 36 654 42 779 50 115 58 929 71 001 120 804 43 209

97 416 113 175 122 897 129 816 137 569 144 851 156 037 168 986 199 058 359 320 155 483

23.05 27.96 28.91 29.31 33.21 38.12 40.25 41.41 45.40 48.69 33.85

7.34 7.25 11.67 13.75 15.15 15.43 16.75 16.17 16.14 10.63 12.81

7.45 6.78 6.88 6.92 6.52 5.91 5.67 5.51 4.95 4.24 6.18

Notes 1 The sample size is 9908 households. Details on the methods used in computing these measures can be found in Wood (1995). 2 Household income inclusive of net imputed rent.

3.21 percentage point reduction in user cost, their annual user cost would fall from 56.9% ($7257) of household income to 32.4% ($4130) of household income, a sizeable reduction in the effective price of housing. We can explore variation in the effective price of housing across the life cycle by grouping households into age groups, using the age of the head of household. In Wood (1995, Table 6.9) households are classiﬁed into three age groups (less than 45, 45 to 65 years and over 65 years) and three household types (married couples with and without dependants, and single persons). For each household type and every income decile, the user cost of capital declines as we move from younger to older age categories. This reﬂects the lower marginal tax rates of younger households.

Investors and rental housing The variation in the user cost of homeowners’ housing capital due to taxation measures is well known. Less well appreciated are the systematic variations in landlord user costs of housing capital due to the tax treatment of their returns. Governments commonly grant tax shelter beneﬁts to investors in rental housing to promote supply. In the USA accelerated depreciation and

116

Housing Economics and Public Policy

the low-income housing tax credits are examples (Hendershott & Ling 1984; McClure 2000). In Australia there has been a building write-off allowance since 1987, and passive losses can be deducted from other sources of income. In the UK various tax-efﬁcient vehicles for the promotion of investment in private rental housing have been created (Wood & Kemp 2001). Capital cost allowances are used in Canada to stimulate the construction of rental housing (MacNevin 1997). In all of these countries there is preferential treatment of real estate capital gains as compared to ordinary sources of income such as rents. In the USA typical project models have been calibrated to explore the effectiveness of tax shelter beneﬁts in lowering the user costs of investors. The approach uses measures of cash ﬂows for a representative rental housing development, to solve for the minimum rental rate at which an investor can meet his/her user cost of capital (see, for example, De Leeuw & Ozanne 1981; Brueggeman et al. 1982; Hendershott & Ling 1984; Fisher & Lentz 1986; Follain et al. 1987; Hendershott et al. 1987; Ling 1992; MacNevin 1997). These studies generally assume that marginal landlords come from the top marginal income tax bracket; in competitive housing markets tax shelter beneﬁts are then passed on to tenants in lower rents as determined by the marginal investors’ income tax bracket. The assumptions that these studies make about tax brackets of marginal investors are then critical.14 To ﬁx ideas consider the present value function of an investor in rental housing who can deduct mortgage interest, real estate taxes and operating costs from assessable rental income; however, capital gains are tax exempt.15 The present value function can be expressed (Wood & Tu 2001) as: (7.8) where τL is the marginal income tax rate of the landlord. The value of R = r(0)/ p(0) that equates the present value of revenues and costs such that V = 0 is the pre-tax real reservation rental rate the project must earn if it is to be an attractive investment. On evaluating the integral in equation (7.8), factoring q and m, re-arranging and solving for R we obtain (Wood & Tu 2001): (7.9) where the right-hand side is the familiar user cost of capital.16 The assumption that capital gains are tax exempt means that if the expected rate of capital appreciation increases by one percentage point, R can fall by a multiple of one percentage point because of the tax that is avoided on a lower rental

Taxation, Subsidies and Housing Markets

117

rate. With capital gains receiving preferential treatment, investors prefer to receive returns in this form. This preference is stronger the higher the marginal income tax rate of the investor. Thus the investor user cost of capital is inversely related to the marginal income tax rate. Wood & Watson (2001) use the Australian Bureau of Statistics 1993 Rental Investors Survey to measure the marginal income tax rates of investors. They found that of the sample of 2906 individual investors, only 315 belong to the top marginal income tax bracket. Thus it would appear that rental housing supply is price inelastic at the reservation rental rates of landlords who belong to the top bracket. It is not known whether this ﬁnding is due to capital market imperfections (Litzenberger & Sosin 1978), an insufﬁcient number of top bracket investors (Sunley 1987) or institutional impediments to multi-property portfolios (Wood & Tu 2001). Whatever the reason, the implications are profound. In order to attract rental investors from lower tax brackets, market rental rates must rise above the levels necessary to meet the reservation rental rates of investors from the top bracket. The latter become intra-marginal investors earning above normal returns. Only part of their tax savings is passed on to tenants in the form of lower rents. In Table 7.3 the landlords comprising the Wood & Watson (2001) study are arranged into deciles according to our user cost measure, and mean user cost, property value and marginal tax rate are listed in each decile.17 The importance of tax preferences in driving variation in user cost is illustrated by the strong inverse relationship between the investor’s marginal tax rate and his/her user cost. The latter increases from a mean of 4.3% in the lowest decile to a mean value of 12.1% in the highest user cost decile. There is a correspondingly monotonic increase in mean market rental rates from 5.9% in the lowest decile, to 9.0% in the highest decile (see Wood & Watson 2001, footnote 41). These ﬁgures indicate that intra-marginal suppliers in the lowest user cost deciles are making above normal returns, while low tax bracket marginal suppliers in the highest deciles are typically making below normal returns. In the longer run we can anticipate exit of these marginal landlords. This brings us to a particularly worrying aspect of the ﬁndings reported in Table 7.3. The column listing mean property values indicates that marginal suppliers are concentrated in low-income rental housing. Any stock losses due to poor returns will then typically occur in this low value segment of the stock. It is signiﬁcant that Yates & Wulff (2000) ﬁnd that in Australia between 1986 and 1996 there has been a healthy expansion in the total stock of rental housing, yet they observe a contraction in the stock of low-income rental housing in that same period.

118

Housing Economics and Public Policy

Table 7.3 Mean property value (Australian dollars), weighted average effective tax rate and user cost by decile of user cost.1

Decile

Mean property value A$

Weighted effective marginal tax rate %

User cost2 %

1 2 3 4 5 6 7 8 9 10

177 560 143 466 142 873 133 944 122 050 125 003 119 327 104 543 94 147 63 604

48.2 41.8 37.1 33.3 29.9 26.0 20.6 20.2 15.5 13.9

4.25 6.17 6.91 7.48 7.98 8.43 8.91 9.41 10.07 12.05

All

122 654

28.7

8.17

Notes 1

Marginal landlords (those in the highest user cost decile) typically have relatively low effective marginal rates of tax. They are also concentrated at the bottom end of the rental housing market.

2 The user cost estimates are computed with respect to the baseline assumptions that i = 10.0% and q = 4.75%. Source: Wood & Watson (2001, table 2).

Tax arbitrage18 Typically we think of taxation measures lowering the cost of owner occupation relative to renting in higher income tax brackets. For a given residence the market rent payable by a potential occupant is independent of their marginal tax rate. However, because net imputed rents and capital gains are tax exempt, the cost of purchasing the given residence for owner occupation falls as the marginal tax rate of the purchaser increases. From this perspective the tax system subsidises ownership among high-income individuals. But there is a sense in which the tax system is biased in favour of renting (Litzenberger & Sosin 1978; Kiefer 1978, 1980; Titman 1982). Above I argued that competitive housing markets will ensure that tax shelter beneﬁts from investment in rental housing will be passed on in the form of lower market rents that reﬂect the tax bracket of the marginal landlord. In the absence of impediments to investors in rental housing, landlords from the highest

Taxation, Subsidies and Housing Markets

119

income tax bracket can be prepared to invest in rental residences, and offer housing services to low bracket taxpayers at a rent that places home purchase at an after-tax cost disadvantage. It has been shown by Anstie et al. (1983), Gordon et al. (1987), Follain & Ling (1988), Hendershott (1988), and Nordvik (2000) that this tax arbitrage process can be analysed using break-even tax rate models. Given a pre-determined demand for housing services, individuals are assumed to choose the tenure that supplies these housing services at least cost. Once appropriate cost equations for the supply of housing in each tenure have been speciﬁed, the models can be solved for the marginal income tax rate at which an individual is indifferent between owning and renting, and landlords are just willing to let housing. Equation (7.6) above can be interpreted as the maximum rental rate that a potential occupant would be willing to pay before turning to purchased housing.19 On the other hand equation (7.9) can be interpreted as the minimum rental rate (the reservation rental rate) that an investor is willing to accept before realising their real estate investment. On setting equation (7.6) equal to equation (7.9) and solving for τ, we derive the break-even tax rate (τ*) given by20 (7.10)

The break-even tax rate describes the conditions under which there is a mutually advantageous landlord–tenant match, assuming that marginal landlords come from the top tax bracket where the tax rate is τmax. These conditions are that the maximum rental rate a potential occupant is prepared to pay must be equal to (or greater than) the landlord’s reservation rental rate. Potential occupants with tax rates less than τ* will ﬁnd it ﬁnancially advantageous to rent rather than own. Their relatively low tax bracket status means that the effective price they must pay to become owner-occupiers is relatively high, while landlords belonging to the top tax bracket can lease housing to them at rents that are ﬁnancially more attractive. A particularly important implication of this model is its prediction that the demand for rental housing will be particularly strong in the low value segments of the market, and this is where tenant–landlord matches are most likely. The low value end of the market is of course more affordable to low bracket taxpayers with τ < τ*, but in addition their maximum rental rates are greater than those high tax bracket individuals are willing to offer, as equation (7.6) above demonstrates. Provided marginal landlords belong to

120

Housing Economics and Public Policy

the highest tax bracket, mutually advantageous landlord–tenant matchings are then most likely at the bottom end of the market.21 Given the evidence provided in Table 7.3 above we should question the validity of this ﬁnal hypothesis, since it shows that marginal landlords do not belong to the top tax bracket, and marginal landlords belonging to lower tax brackets are concentrated at the bottom end of the market. Table 7.4 presents the results when we calibrate the break-even tax rate model using the Australian Bureau of Statistics 1993 Rental Investors Survey. The 1907 properties are arranged into deciles according to landlords’ reservation rental rates. Mean break-even tax rates are listed for each decile, and for three loan to value ratios (LVR). At the high LVR of 0.75 the mean break-even tax rate is 77.8%. All Australian taxpayers have marginal tax rates less than 78% (see ﬁnal column), which implies that renting is relatively less costly than purchase for all taxpayers at this LVR. Even at the lower LVR of 0.5 the mean break-even rate of 38.9% implies that more than three quarters of Australian taxpayers ﬁnd renting relatively less costly. Only at the low LVR of 0.25 does the relative cost of homeownership become more favourable for a majority of taxpayers. The other notable ﬁnding in Table 7.4 is that the break-even tax rate declines as we move to higher reservation rental rate deciles. Yet these higher deciles contain the concentrations of low value rental housing that is affordable to low-income groups (see Table 7.3). Successful tenant–landlord matchings are least likely in these segments, which contradicts the pattern predicted by the break-even model. This unexpected ﬁnding arises because marginal landlords come from the lower tax brackets and are concentrated in low value segments. If housing and capital markets were competitive the supply of rental housing would be perfectly elastic at the reservation rental rates of top tax bracket investors. As was pointed out earlier, the presence of low tax bracket investors implies that market imperfections of one form or another impede additional rental investments by top bracket taxpayers. It is also perplexing to observe that many Australian households ﬁnd homeownership ﬁnancially unattractive given the predominance of owner occupation in Australia. It would seem that preferential taxation is not responsible for the high rates of home ownership observed in Australia and countries with similar taxation arrangements. The puzzle may be resolved on recognising that certain costs, such as agency costs, are unique to rental housing. Landlords cannot write perfect rental contracts that cover all pos-

Table 7.4 Mean break-even tax rate and % of taxpayers with marginal tax rate (MTR) less than break-even tax rate (BTR) by decile of the landlord’s reservation rental rate (land taxes, transaction costs and agency costs = 0).1 Break-even tax rate % (LVR = 0.5)

Break-even tax rate % (LVR = 0.75)

% of Taxpayers with MTR < BTR (LVR = 0.25)2

% of Taxpayers with MTR < BTR (LVR = 0.5)2

% of Taxpayers with MTR < BTR (LVR = 0.75) 2

1 2 3 4 5 6 7 8 9 10

59.57 40.75 34.17 29.42 24.11 20.96 16.58 13.11 10.94 9.76

89.36 61.12 51.25 44.14 36.17 31.44 24.86 19.66 16.41 14.64

178.7 122.2 102.5 88.27 72.34 62.88 49.73 39.32 32.83 29.29

100 79.1 41.6 41.6 41.6 41.6 0.55 0.55 0.55 0.55

100 100 100 79.1 41.6 41.6 41.6 0.55 0.55 0.55

100 100 100 100 100 100 100 79.1 41.6 41.6

All

25.93

38.89

77.79

41.73

79.23

1003

Notes 1 The break-even tax rates are computed for the baseline case in which the expected rate of inﬂation is 4.75%. 2 The Medicare levy is ignored for this calculation. LVR is loan-value ratio. 3 At the sample mean break-even tax rate. Source: Wood (2001), Tables 1A and 1B.

Taxation, Subsidies and Housing Markets

Decile

Break-even tax rate % (LVR = 0.25)

121

122

Housing Economics and Public Policy

sible contingencies, thus tenants do not face the social marginal cost of their utilisation of the rental housing stock (‘the fundamental rental externality’; see Henderson & Ioannides 1983). Agency costs are the resources devoted to screening tenants, negotiating contracts, compiling inventories, conducting property inspections, collecting rents and whatever monitoring or policing actions are called for in managing the rental externality associated with landlord–tenant relationships. We obtain a proxy measure for agency costs by setting them equal to the property management and letting fees landlords pay real estate agents, if the latter are engaged to undertake all monitoring and policing tasks.22 Our measure of agency costs is incorporated into the break-even model and the results are presented in Table 7.5.23 We now ﬁnd lower break-even tax rates at all LVR employed in the microsimulations. At the mean break-even tax rates for these LVR only 1% of taxpayers ﬁnd mutually advantageous tenant–landlord matches. Thus home ownership has a relative cost advantage over renting for most individuals, but taxation factors are not the only explanation. The agency costs that landlords incur as a result of managing the principal/agent relationship are avoided when people rent from themselves as owner-occupiers. This is an important source of owner occupation’s relative cost advantage. A striking ﬁnding is the negative break-even rates found in the ﬁve highest reservation rental rate deciles. This indicates that mutually advantageous tenant–landlord matches cannot be found, unless there are direct rental subsidies that overcome homeownership’s relative cost advantage. It is in these segments that the break-even tax model predicted that landlord–tenant matches would be most common. Critically important here is the observation that marginal landlords in these high reservation rental rate deciles are concentrated in low value segments of the rental market. One possible explanation for this ﬁnding is rent clientele effects, an inefﬁciency in housing market processes that I address in the next section.

Table 7.5 Mean break-even tax rate and % of taxpayers with marginal tax rate (MTR) less than break-even tax rate (BTR) by decile of the landlord’s reservation rental rate (inclusive of land taxes and agency costs).1 Break-even tax rate % (LVR = 0.5)

Break-even tax rate % (LVR = 0.75)

% of taxpayers with MTR < BTR (LVR = 0.25)2

% of taxpayers with MTR < BTR (LVR = 0.5)2

% of taxpayers with MTR < BTR (LVR = 0.75)2

1 2 3 4 5 6 7 8 9 10

46.25 24.48 15.18 9.06 3.58 –1.89 –6.06 –8.71 –13.41 –14.81

69.38 36.72 22.77 13.58 5.37 –2.84 –9.09 –13.07 –20.11 –22.11

138.76 73.45 45.53 27.17 10.73 –5.67 –18.17 –26.14 –40.23 –44.42

79.1 41.6 0.55 0.55 0.55 0 0 0 0 0

100 41.6 41.6 0.55 0.55 0 0 0 0 0

100 100 79.1 41.6 0.55 0 0 0 0 0

All

5.36

8.04

16.07

0.553

0.553

Notes 1 The break-even tax rates are computed for the baseline case in which the expected rate of inﬂation is 4.75%. 2 The Medicare levy is ignored for this calculation. LVR is loan-value ratio. 3 At the sample mean break-even tax rate. Source Wood (2001), Tables 2A and 2B.

0.553

Taxation, Subsidies and Housing Markets

Decile

Break-even tax rate % (LVR = 0.25)

123

124

Housing Economics and Public Policy

Taxation, inefﬁciency and market processes Clientele effects In the previous section I argued that there appear to be impediments that prevent additional investment by top tax bracket investors in private rental housing. As a consequence lower tax bracket investors, who have higher costs, are attracted into the market by the higher rents required if they are to cover those inﬂated costs. The costs of supply are therefore higher than would otherwise be the case. The consequences of this inefﬁciency are particularly severe in low-income rental housing because marginal investors are concentrated in this low value segment. An explanation for this pattern is important if we are to address the issue from a policy perspective.24 Wood & Tu (2001) develop a housing segmentation model in which there are two sub-markets X and Y. They make the following assumptions.

•

In a market offering rental housing services only, housing is identical in all respects except location. Investors’ user costs of capital are as described in equation (7.9).

•

The expected rate of property price appreciation is higher in X than Y.

•

There are two income tax brackets with marginal rates higher in the top bracket.

•

There are capital market imperfections, which are such that in the short run high bracket investors can ﬁnance acquisition of rental housing in sub-market X, but they are constrained on attempting to ﬁnance the acquisition of all housing in Y.

•

Rent levels are sticky in the short run and the housing stock is ﬁxed in the short run.

•

Capital and rent values are initially the same and the former are equal to replacement costs.

High bracket investors will compete with low bracket investors to enter submarket X, but the former gain more from tax exempt capital gains and will be prepared to pay a relatively high premium to acquire rental investments in X. Market values in X begin to rise, while in Y market values fall in order

Taxation, Subsidies and Housing Markets

125

to attract investors through higher market rental rates. Equilibrium requires satisfaction of two arbitrage conditions. Firstly, equalisation of post-tax real returns for high bracket investors ensures that they are indifferent between X and Y. Secondly, market rental rates in Y must rise such that low bracket investors ﬁnd it just proﬁtable to invest in Y. Wood & Tu (2001) demonstrate that when there is simultaneous satisfaction of these two arbitrage conditions high bracket investors become intra-marginal suppliers while low bracket investors are marginal suppliers in sub-market Y. The latter are deterred from investing in X by the prospect of negative returns. The model’s principal proposition is that we can expect to observe tax-driven investor clienteles in which low (high) tax bracket investors are concentrated in low (high) value segments. This is consistent with the evidence presented in Table 7.3, and demonstrates that there can be supply-side sources of housing market segmentation where market rental rates for identical housing are driven apart by the clustering of tax-driven investor clientele groups. Wood & Tu (2001) estimate a two-equation model using cross-section data on market rental rates and determinants from the Australian Bureau of Statistics 1993 Rental Investors’ Survey. The degree to which capital gains are given tax preferential treatment is found to have a statistically signiﬁcant negative impact on market rental rates, which is consistent with the operation of a clientele effect. There is also an interesting corollary of this model. In the short run, Tobin’s-q will exceed one in the high value segment X, so that we can expect expansion in the higher value stock in the long run. On the other hand, Tobin’s-q will be less than one in the low value segment Y and we can expect contraction in this segment.25 Some part of the ﬂow of resources into segment X would occur even if there were symmetric tax treatment of capital gains and rental income. But the privileged tax status of capital gains ampliﬁes the ﬂow of resources into segment X, and is the source of market inefﬁciency. One policy response is the introduction of tax preferences targeted on landlords holding low-income rental housing (see Wood et al. 2001 and the concluding section below).

Lock-in effects Capital gains taxation measures can be the source of lock-in effects that cause investors to hold on to rental housing assets and discourage changes in portfolios. The effect arises because of the failure to tax gains as they accrue. By deferring realisation of an investment the present value of a capital gains tax liability is reduced. As a consequence investors end up holding a

126

Housing Economics and Public Policy

less preferred set of investments and welfare losses ensue. Gravelle (1994, p. 143) considers these lock-in effects to be the primary argument for reducing capital gains tax if the effect is a powerful one. This is because large welfare gains will be generated relative to revenues forgone, and it is even possible that lowering the tax rate would actually raise revenues through increased realisations. The magnitude of lock-in effects will reﬂect the way in which capital gains taxes are applied. For instance the effect will be magniﬁed if: • the tax is forgiven on death, as in the USA; • depreciation allowances are recaptured by subtracting allowances from the cost base used to calculate taxable capital gains, as in Australia and the USA; • capital gains are taxed by the ‘top slice’ method that applies to increments in ordinary income, as in the UK. On the other hand churning incentives will reduce the magnitude of lockin effects (Hendershott & Ling 1984; Gordon et al. 1987; Gravelle 1994). Churning incentives are evident when the purchaser of a building can begin depreciation again, based on the sale price of the asset, as in the USA. If the difference in the present value of the tax shelter beneﬁts of depreciation in the hands of the seller and the hands of the purchaser exceed the capital gains tax, then realisation is tax advantaged (Gravelle 1994, p. 135). However, there are other economic factors that can impact on the size of lock-in effects. To ﬁx ideas consider a real estate investment that has been held for T years. If the investment is realised now, capital gains tax liabilities are given by: λτL q[p(T) – p(0)]

(7.11)

where 0 < λ < 1 is the tax preference with respect to capital gains tax, and all other parameters are as deﬁned above. If the landlord continues to hold the asset for a further n years, the present value of capital gains tax liabilities on gains accrued up to year T are: λτL q[p(T) – p(0)]e-kn

(7.12)

Taxation, Subsidies and Housing Markets

127

On subtracting (7.12) from (7.11) we obtain a measure of the value of deferring capital gains tax liabilities. This is given by: λτL q[p(T) – p(0)](1 – e-kn)

(7.13)

It is evident from equation (7.13) that the option to defer has a value that is increasing in the rate of capital appreciation, the number of years the investor has held the asset, the expected number of years that the investor expects to continue to hold the asset if the option to defer is chosen, and the rate of inﬂation (and hence discount rate). We can therefore expect lock-in effects to be particularly signiﬁcant in inﬂationary conditions that are accompanied by buoyant property prices. In addition, lock-in effects are typically larger for older investors, because they are likely to have held assets for a longer period. When an investor is considering switching into a new asset, the rate of return which the new asset yields has to increase by an amount at least equal to the value in equation (7.13), if it is to leave the investor as well-off as holding the existing asset. This approach to the measurement of lock-in effects was originally suggested by Holt & Shelton (1962). Gravelle (1994, Table 6.6) has applied the measure to building structures, but there have not been other applications to real estate where the focus has been on churning incentives. Similarly there have been no attempts to measure the welfare losses from lock-in effects with respect to real estate assets. Typically studies have concentrated on corporate stock (Feldstein et al. 1980) or the use of tax return databases where no distinction is drawn between assets (Auten & Clotfelter 1982). This is clearly an area worthy of research.26

The demand for housing, over-investment and welfare losses Clientele and lock-in effects are a relatively recent concern in the real estate literature concerning taxation and efﬁciency losses. The early literature focused on measurement of the deadweight welfare losses consequent upon using tax expenditures to promote the consumption of owner-occupied housing. In an early study of this kind Laidler (1969) assumes a perfectly elastic long-run supply of housing and a –1.5 price elasticity of demand. Using an approach similar to equation (7.7) above, Laidler calculates the impact of tax subsidies on price, and estimates that the stock of housing in the USA would be 17.1% lower if imputed rent had been taxed. The excess burden or welfare loss triangle corresponding to this subsidised increase in

128

Housing Economics and Public Policy

consumption sums to only 0.14 of 1% of the value of the housing stock in the USA. King (1981) estimates that in the UK taxation of imputed rent would result in a 13.7% decline in the long-run consumption of housing services, a ﬁgure very similar to that obtained by Laidler (1969). Once again the excess burden is small at only 0.4 of 1% of mean income, and those with the highest incomes are worst hit. Rosen (1979) uses a cross-section database to estimate jointly the quantity of housing services demanded and tenure choice. If taxation of imputed rent was in operation, this study projects that the incidence of owner occupation would be 4.4% lower. Rosen & Rosen (1980) estimate a time series model of the tenure choice decision for the period 1949–1974. Imputing measures of the relative price term in the absence of tax preferences allows the authors to employ their regression model for the purposes of generating ex-post forecasts of the long-run proportion of home owners. Their regression model indicates that the incidence of owning would fall from 64% to 60%. Using the homeownership rate adjusted for changes in the demographic structure of the population, Hendershott & Shilling (1982) project that the incidence of homeownership would be 59% if property taxes and mortgage interest were not deductible. The positive impacts of tax subsidies on owner-occupied housing investment that are implied by the above studies raise the possibility that tax induced increases in housing investment have crowded out industrial capital. When capital gains are tax exempt and mortgage interest is deductible, accelerating inﬂation raises the cost of investment in industrial capital relative to owner-occupied housing (Rosen 1985). Simulation studies based on the typical project approach incorporate this relative cost into investment functions, and show that when inﬂation increases owner-occupied housing investment rises relative to business investment (Hendershott & Hu 1981). In Mills (1987) a two-sector neoclassical growth model is estimated for the US economy, and employed to estimate the efﬁcient allocation of capital stock. The model estimates imply that the housing stock is 33% higher than is optimal. Wood’s (1988a) two-sector neoclassical model permits estimation of inter-sectoral externalities and factor input productivity differentials. When estimated for the Australian economy the model estimates indicate that housing is responsible for the generation of positive externalities that promote growth of the economy. In addition factor inputs appear to be no less productive when deployed in the housing sector rather than in the rest of the economy.27

Taxation, Subsidies and Housing Markets

129

These conﬂicting results suggest the need for research that identiﬁes the existence and signiﬁcance of the transmission mechanisms through which positive and negative externalities manifest themselves. The more recent literature has been exploring the links between homeownership on the one hand and health, neighbourhood stability, social involvement and socially desirable behaviours on the other.28 These examples of potentially positive externality transmission mechanisms are being balanced by a recent spurt of studies into the potentially adverse impacts of an individual’s home ownership on the duration of unemployment spells, and the link between regional unemployment rates and the incidence of homeownership (Oswald 1997b; Green & Hendershott 2001a).

Concluding comments A number of important policy concerns are evident from this overview of taxation and housing markets. In respect of owner occupation a consistent issue raised by commentators has been the inequitable distribution of tax subsidies. Given typical taxation arrangements it has been ﬁrmly established that the effective price paid by homeowners varies inversely with their marginal income tax rate. In addition younger low-income homeowners with low net worth face a relatively high effective price. The belief that typical taxation arrangements encourage expansion of homeownership at the expense of rental housing and even industrial capital, and is therefore responsible for allocative inefﬁciencies, has not attracted so much concern in recent years. This is because of a growing number of studies offering empirical support for the idea that encouraging owner occupation yields signiﬁcant social beneﬁts (Vandell 2000). In addition, there is the view that tax shelter beneﬁts available to high tax bracket landlords are passed on in lower rents, that advantage households with zero or low marginal tax rates (Weicher 2000). Policy responses to the concerns about the distribution of homeowner tax preferences initially focused on introducing taxation of net imputed rents. However, it has been shown that the resulting increase in tax liabilities can be regressive, with elderly ‘asset rich–income poor’ households particularly adversely affected (Ling & McGill 1992; Bourassa & Hendershott 1994). Transitional arrangements can be framed to mitigate these concerns, and a ‘package’ of reforms can be assembled whereby taxation of net imputed rents replaces regressive real estate taxes and transaction taxes, thereby

130

Housing Economics and Public Policy

ensuring a progressive redistribution of the tax burden (Wood 1995). But it seems that most policy analysts consider taxation of net imputed rents a politically unrealistic policy reform. Less ambitious measures such as the phased removal of mortgage interest and real estate tax deductions (Bourassa & Grigsby 2000), or the conversion of deductions into tax credits (Green & Vandell 1999) are now seen as more realistic measures. As noted earlier, there is a change in sentiment with respect to homeowner tax concessions, in view of the growing number of studies conﬁrming social beneﬁts from expansion of homeownership. These studies strengthen the case of those who favour retaining tax concessions, but in a form that would extend the beneﬁts further down the income distribution. A more recent issue has emerged as a result of evidence conﬁrming a growing shortage of low-income rental housing (Malpezzi & Green 1996; Yates & Wulff 2000; Somerville & Holmes 2001), and the role of taxation measures as one of the factors causing this phenomenon (Wood & Tu 2001). It appears as though there are impediments that cause the supply of rental housing to be price inelastic at the reservation rental rates of investors belonging to the top tax bracket. As a consequence market rental rates rise in order to attract low bracket investors, and only part of the tax shelter beneﬁts received by top bracket investors are passed on in the form of lower rents. To the extent that marginal (low bracket) landlords are concentrated in low value segments of the rental housing market, this has particularly severe impacts on affordable rental housing for low-income groups. It turns out that successful tenant– landlord matches are then least likely in the low value segments, and the exit of low-income housing landlords is to be expected. The uncompetitive position of this segment of the private rental housing stock is compounded by relatively high agency and transaction costs for landlords investing in such rental housing (Wood &Watson 2001). A number of potential policy responses to this issue deserve attention. The ﬁrst focuses on relaxing any impediments that deter residential rental investment by high tax bracket investors. These impediments could arise due to capital market imperfections, government regulation or ﬁscal intervention. There is a paucity of evidence here that could guide policy formulation. A second approach emphasises the use of targeted tax subsidies to sharpen the incentive to invest in low-income rental housing. In the USA the 1986 Tax Reform Act reduced the tax shelter beneﬁts available to rental housing investors, and raised concerns that low-income rental housing investment would decline. A targeted low-income tax credit was introduced to allay these fears, but its administrative complexity and limited scope has caused analysts to question its effectiveness (McClure 2000). However, Wood et al.

Taxation, Subsidies and Housing Markets

131

(2001) show that a low-income tax credit targeted at rental units that are in the lowest quartile by rent, and which is open-ended, lowers the mean KingFullerton effective marginal tax rate from 61% to 37%. A third approach is based on the observation that agency costs are relatively high in the low value segment of the private rental housing stock. It is in this segment that we are most likely to observe marginally housed tenants, whose attachment to the labour market is weak and where rent arrears, unanticipated breaches of lease contracts and other agency problems are most severe. Vertical disintegration, in which privately ﬁnanced rental housing is managed by social housing agencies who have professional expertise in housing management, could lower agency costs sufﬁciently to restore the competitive position of this segment of the rental housing stock. Given signs that access to home ownership is increasingly difﬁcult, particularly for those on low incomes, a strong demand for rental housing that is affordable to low-income groups can be anticipated. Evidence that this segment of rental housing is costly to supply, and contracting in size, is then a serious concern warranting a policy response.

Acknowledgements The author is grateful for the constructive and helpful comments provided by Christine Whitehead, Kenneth Gibb, Tony O’Sullivan and participants at the workshop held at Glasgow University, 9th and 10th November 2001.

Notes (1)

(2) (3)

The analysis presented in this chapter assumes an Anglo-American ﬁscal regime in which there is a relatively high reliance on direct taxes levied on income. In other countries such as those in Southern Europe, there is a greater reliance on indirect taxes. The focus on Anglo-American ﬁscal regimes reﬂects a bias in the English language literature, and the author’s empirical work that has been conducted using Australian and British data. An exception is the US low-income housing tax credit introduced in 1987 (McClure 2000). Though the early literature emphasised static deadweight welfare losses, more recently there has been an increasing recognition of the positive externalities associated with increases in the size and quality of the housing stock. If sufﬁciently large these positive externalities can justify a ﬁscal regime that discriminates in favour of housing. I consider this further later in this chapter.

132 (4) (5)

(6)

(7) (8)

(9)

(10)

(11) (12)

(13)

(14)

Housing Economics and Public Policy These concessions are also enjoyed by owners of second homes that are not rented out (Bourassa & Grigsby 2000, p. 524). Thus the notion of equity that I have in mind is that all households pay the same price for a given bundle of housing services. I should acknowledge an alternative view of equity. In a progressive tax system tax burdens are relatively onerous for high-income households. If governments decide to lower the tax burden by introducing tax expenditures, higher income households will beneﬁt more from the reduction in the tax burden. This is ‘fair’ because it ensures symmetry in the tax system, given that higher income households will contribute proportionately more of their income in tax. Deﬁned to include both land and the building structure. The ratio of land to building structures is assumed to be ﬁxed. The capital variable could comprise land and building structures from one or more properties. Economic depreciation is ignored. The ﬁrst two terms on the right-hand side of equation (7.1) are the homeowner’s initial equity stake and discounted sales proceeds. The third term is the present value of capital gains tax liabilities when gains are taxed as they accrue and at the homeowner’s marginal income tax rate. The ﬁnal term is the present value of after-tax net imputed rents. Transaction costs have been ignored. Or as Chinloy (1991, p. 516) deﬁnes the term, it is ‘the price which an owner occupier must pay to obtain a unit of services, while owning a unit of stock’. The chosen benchmark is the tax treatment prevailing in Australia, Canada and New Zealand, where imputed rents and capital gains are tax exempt, but mortgage interest and real estate taxes cannot be deducted. Now that the mortgage interest deduction has been phased out in the UK, the present value function will also be the same in that country as well. Under the tax neutral provisions this term disappears because the real value of the interest deduction is also eroded. At higher marginal income tax rates the tax beneﬁt from non-taxation of the erosion of the real value of mortgage debt becomes more important relative to the tax penalty associated with debt. Compare the second and fourth components of equation (7.6). This method involves calculation of the homeowners’ tax liability inclusive and exclusive of their net imputed income. The resulting incremental tax liabilities are expressed as a proportion of net imputed income. The Australian statutory marginal tax rate schedule in 1990 was applied in this calculation. The marginal tax rate estimates obtained by this method can exceed the highest statutory marginal tax rate because we have taken forgone tax rebates into account. Homeowners with a high-income elasticity of demand will, ceteris paribus, have larger net imputed rents and higher estimated marginal tax rates. It follows that homeowners with a high-income elasticity of demand will tend to have a lower user cost of capital. Narwold (1992) has offered a theoretical rationale for the top marginal income tax bracket assumption.

Taxation, Subsidies and Housing Markets (15)

(16)

(17)

(18) (19)

(20)

(21)

(22)

(23)

(24)

(25)

133

The role of inheritance taxes has been ignored. The evidence presented below uses Australian data. In this country estate and inheritance taxes were abolished by all state governments by the early 1980s (see Wood 1992). If the investor cannot access depreciation allowances, and capital gains are tax exempt, this user cost expression is the same as that applicable in the USA. In Hendershott & Slemrod (1983), Gordon et al. (1987), Follain & Ling (1988) and Narwold (1992) the user cost expression reduces to equation (7.9) under these conditions. The user cost measure estimated in Wood & Watson (2001) is in fact more comprehensive than that expressed in equation (7.9). Measurement of landlords’ marginal income tax rates uses similar methods to those used in measurement of homeowner marginal income tax rates (see Note 14). This section is based on Wood (2001). Equation (7.6) is the user cost of capital for homeowners under typical taxation arrangements. If the individual’s pre-determined demand for housing can be met at a lower market rental rate, renting is the ﬁnancially more attractive option. In the derivation of equation (7.10) the component in the homeowner’s user cost expression representing erosion of the real value of mortgage debt has been omitted to ensure a unique solution. The consequences of its omission are minor in a low inﬂation environment, as has been typical of the 1990s. The consequences are even less important for individuals in low tax brackets, who are a particular focus of concern here. In the absence of rental subsidies, the top end of the market is affordable to high bracket taxpayers only. But given τ > τ* the maximum rental rates they are willing to offer are lower than those which low bracket individuals are prepared to offer for affordable residences. A landlord–tenant matching is then much less likely at the top end of the housing market. A wider range of costs than this is relevant. For example, there are rent arrears, vacancies and delays in the payment of direct housing subsidies that are formally transferred to landlords by government agencies on behalf of tenants. Thus the proxy measure is imperfect. I am at present engaged in a programme of research that is seeking to provide more precise and inclusive measures of these costs in different value segments of the housing stock. See equations 4 and 5 of Wood (2001) for a derivation of the agency cost version of the break-even model. Land taxes are also included in this version of the model. Also critical is identiﬁcation of the impediments that prevent expansion of the supply of private rental housing by top bracket investors. This question is outside the scope of this chapter, but is an important item on a future research agenda. In addition the model implies that current market rental rates should predict future rates of price appreciation. Empirical support for this hypothesis can be found in Capozza & Seguin (1995) and Clark (1995).

134 (26)

(27)

(28)

Housing Economics and Public Policy Real estate has an added signiﬁcance in this respect since transaction costs are generally thought to be large, and will be an additional source of lockin effects. Indeed real estate’s illiquidity can be traced to lock-in effects. Real estate illiquidity is thought to be the source of rent premiums that vary according to investor holding periods. Liu et al. (1995) argue that these premiums give rise to ‘dividend-like’ clientele effects, and Wood & Tu (2001) offer evidence to suggest that these premiums are a further source of supplyside segmentation. These two-sector growth models draw on the production function approach to changes in aggregate output deployed by Feder (1982) and Ram (1986) to investigate the relationship between size of the export sector, government sector and economic growth. See Rohe et al. (2000) and McCarthy et al. (2001) for a review of this literature.

8 The Economics of Social Housing Christine M. E. Whitehead

Introduction: the role of economic analysis in social housing Some thirty years ago when Duncan Maclennan was ﬁrst beginning to work in housing economics the concept of social housing was a simple one. In the UK, the objective of housing policy was seen to be ‘a separate house for every family that wishes to have one’ (Ministry of Reconstruction 1945). The need was for as much housing as possible – anywhere and of any type – as a result not only of wartime destruction and massive under-investment during and immediately after the war but also because of the rapid growth in population and households in the post-war era. The UK government approach to meeting this need was through public ownership and development of large numbers of dwellings, rents set at well below market levels and administrative allocation of the available public sector housing. Public sector provision was seen as the only way forward, not simply on ideological grounds, but also because of limited private sector capacity, the fact that much of the land available for housing was in public hands, and that the slum clearance and redevelopment of urban areas necessary to meet appropriate standards required large-scale direct local authority involvement. As a result the 1960s in particular had seen the longest period of sustained expansion in public housing output ever experienced in the UK. The picture was very similar in much of Europe as a result of massive shortages. Policy was simply a numbers game and public production the easiest means of alleviating that shortage. Equally there had been very little economic analysis of either the rationale of social housing or the impact of particular policies on housing outcomes. The requirement of a separate home was seen to be obvious, as was the need for subsidy to public provision and the beneﬁts of pooling costs across the existing stock and new development of social housing. Instead, economic analysis in the UK tended to be limited to examining positive economic

136

Housing Economics and Public Policy

questions relating to housing and the macroeconomy and to changes in incomes, prices and costs (Whitehead 1974). Yet the picture was beginning to change. Most importantly, the 1971 Census for the ﬁrst time showed that the number of dwellings exceeded the number of households in almost every local authority area in the UK. The 1971 White Paper ‘Fair Deal for Housing’ no longer looked simply to numbers, but spelled out a more complex set of objectives deﬁned as ‘a decent home for every family at a price within their means’ (Department of the Environment 1971), raising issues of quality and price for the ﬁrst time. The idea that some dwellings might already be unacceptable led the Department to undertake a survey of difﬁcult-to-let properties – the response to which showed examples in almost all authorities. Nor were demolition and replacement any longer seen as a panacea for the problem of slums. Instead the emphasis moved towards rehabilitation and area improvement – recognising that the idea of a decent home was no longer restricted to the dwelling itself but included the surrounding area (Paris & Blackaby 1979). It was in this context that Anthony Crosland, the then Secretary of State for the Environment, ordered a full review of housing policy in England, and economics for the ﬁrst time played a more central role in policy analysis. This marked the beginning of a far more analytic approach to policy evaluation (Department of the Environment 1977). The link between economic analysis and policy development is not simply about understanding what is happening. If it is to be of any value there must be some capacity to modify government policy as a result of that analysis. It is in this context that Duncan Maclennan has made his most immediate contribution, not just through detailed studies of how policy has operated and how it could be improved but in his preparedness to take a central role in arguing for policy change consistent with these analyses. Policy change over the last twenty-ﬁve years has inherently been incremental, and often seems to have been aimed as much at undoing the problems arising from earlier policies as at developing a proactive approach to changing requirements. Economic analysis, on the other hand, has concentrated not only on assessing particular measures, but also on examining more fundamental issues underlying these policies. Even though politicians are usually pragmatic and better at adding policies than at rationalisation, there is a coherent economic framework that can be applied to the analysis of government intervention in housing and against which policies can be assessed.

The Economics of Social Housing

137

In the context of social housing these more fundamental issues include the following.

•

Does meeting government objectives necessarily imply a signiﬁcant social housing sector? Other countries with equivalent commitment to adequate accommodation for all, notably Germany and Switzerland (as well as the USA and Australia where the commitment can be argued to be less), have effectively met social objectives through private provision.

•

Does the existence of a social sector necessarily imply public ownership? Again, clearly not in many European countries such as the Netherlands, France and Sweden, where independent social landlords play a major, sometimes core, role in social provision.

•

Does the existence of a large social sector necessarily imply below-market rents? Again there are examples, such as in Sweden and the Netherlands, where rents in the social sector are as high or higher than in the market sector.

•

Should there be administrative allocation of social housing to ensure that certain groups obtain adequate accommodation, or should there be consumer choice between sectors? Again many European countries, while giving priority to particular groups, provide tenure neutral assistance, so that social objectives can be achieved across the whole stock (Turner & Whitehead 1993; Turner et al. 1996; Stephens et al. 2002).

These are fundamental policy issues but they are also issues where economic principles can make a relevant contribution. In particular economics can address:

•

whether markets can provide housing efﬁciently as long as people have adequate purchasing power – in which case there may be no need for a speciﬁc social housing sector;

•

whether there are economic reasons why housing itself should play a proactive role in income redistribution through supply rather than demand subsidies together with administrative allocations to ensure that priority needs are met; and

•

whether there are reasons why the social governance of housing might be more cost-effective than private provision – which would suggest a reason for public or at least social ownership to address questions of risk bearing, management and funding.

138

Housing Economics and Public Policy

This chapter concentrates on these three issues by examining the efﬁciency rationale for government intervention, the use of housing as a means of redistribution, and the relevance of transactions costs and ‘new’ institutional economics to the appropriate governance structure for housing provision and management. Addressing these issues helps to provide some insights into the appropriate role for social housing in the twenty-ﬁrst century.

Efﬁciency reasons for social housing provision Mainstream economics provides a well-deﬁned typology of the rationale for government intervention to achieve allocative and productive efﬁciency in a market system. This concentrates on the situations where prices do not properly reﬂect social values and where there are factors that inhibit effective adjustment to price incentives. It also suggests that government may intervene in three fundamental ways – regulation, both to deﬁne the framework within which individual decisions occur and directly to control price or quantity; taxation or subsidy; and direct provision of relevant goods and services.

Sources of market failure Housing has many of the characteristics likely to generate market failure and therefore both productive and allocative inefﬁciency. Equally, all types of government intervention have been applied to its provision, not only in the UK, but across the industrialised world (Charles 1977; Maclennan 1982; Whitehead 1984). Most fundamentally, housing lies somewhere along the spectrum between private goods, where the market can be expected to work well, and public goods – where the market will generally not provide effectively (Barr 1998). In this context, housing is very clearly a private good, in that the consumption of housing is both rival and excludable and the vast majority of beneﬁts go to the owner or the occupier of the dwelling. Most importantly a house provides very different utility to different people – a ﬂat in the centre of town may be exactly what a young couple wants, while a family spending the same amount of money is likely to choose a house, perhaps with a garden, further out of town. It is this emphasis on the value of choice that lies at the core of libertarian views that housing be privately provided and allocated, with assistance concentrated on income support (Whitehead 1991).

The Economics of Social Housing

139

However, the housing market does suffer from signiﬁcant market failures, at least at the bottom end of the market – and it is these that generate the economic rationale both for intervention and social provision. The most important sources of such failure are seen as arising from externalities of three types: direct, interactive and intergenerational. The majority of direct externalities relate to health in that poor housing can encourage the spread of disease, although generally only at standards well below those prevalent in advanced economies (Burns & Grebler 1977). In this context much of the emphasis has been on a government regulatory framework that applies to all sectors. This helps determine location, density, design, building materials and occupation levels. Social housing provision still has a role because these standards may not be affordable to all, and regulations are difﬁcult to enforce – so direct provision can provide an efﬁcient approach to ensuring compliance. An important current example here is the evidence that many of the most distressed areas in England are made up of private rented housing, which is in far worse condition than the social housing nearby (Smith 1999). More generally, some 30% of unﬁt housing actually in use is in the private rented sector (Department of Environment, Transport and the Regions 1998). Interactive externalities occur across the whole spectrum of provision, wherever the use or upkeep of a particular property impacts on the value of another, and vice versa. Thus if I paint my property some of the beneﬁts fall to my next-door neighbours, while if I do not bother, some of the costs are also borne by them – resulting in an incentive to under-maintain and perhaps even to generate slums (Davis & Whinston 1961; Rothenberg 1967). Social housing addresses this issue by internalising the externalities, but whether the incentives are there for optimal investment in maintenance and improvement is far from clear. Externality inefﬁciencies are also the major source of more general neighbourhood decline, where there is no incentive for individuals to invest up to the appropriate level across local services and neighbourhood management, nor a capacity to enforce private contracts which attempt to overcome these failures (Davis & Whinston 1961; Galster & Killen 1995). The problems apply just as strongly to the land market, where externalities, both positive and negative, generate a strong case for land use planning and use allocation to ensure optimal locational patterns (Harrison 1977). Intergenerational externalities arise both because ﬁnancial markets are themselves imperfect and because the social discount rate is lower than the private rate – individuals die, while society continues into the future. As a result, individual decisions are likely to emphasise more short-term and

140

Housing Economics and Public Policy

consumer-oriented expenditures while society, as a whole, requires longerterm solutions (Hirshleifer 1970). This suggests that, left alone, the private market will under-invest in both new development and improvement investment. Social housing is a direct way of generating both higher levels of investment and more fundamental improvement responses. These problems are exacerbated by both lack of information and asymmetry in that information (Macho-Stadler & Pérez-Castrillo 1997). This means that consumers cannot effectively evaluate the beneﬁts of their own investments nor choose appropriate suppliers. In particular the housing market does not properly reﬂect the utility of these investments, so that the private owners cannot readily realise their value nor borrow to obtain the required ﬁnance. More generally, individuals tend to be risk averse, which again reduces the incentive to invest and supports the case for government intervention to increase the resources for housing investment – in that government decisions should be based on risk neutrality (Arrow & Lind 1970; Arrow 1971). A further reason for government involvement relates to problems of monopoly power. Traditional forms of market power are not inherently a major failure of housing markets. In particular, economies of scale either in production or management appear to be relatively limited, while in most areas there are large numbers of housing suppliers. There are, however, signiﬁcant problems associated with interpersonal contracts, particularly in the context of landlord and tenant relationships where there is often asymmetric information, and contracts are not based on equal access to information. These can be addressed either by social ownership, where the objective is to maximise tenant beneﬁts and social value, or by owner occupation, which internalises the contractual arrangements. On the other hand social ownership can generate large-scale administrative failure, if objectives are poorly speciﬁed and agency relationships inadequately addressed. In particular social ownership brings with it its own monopoly power in localities where it becomes the main source of lower-income rented housing. One of the most prevalent market failures relates to the problems of slow adjustment in the housing system – and most importantly the fact that in market systems price can adjust far more rapidly than quantity. Thus in times of sudden reduction in supply, increases in demand or rapid migration, markets will generate adverse distributional outcomes, some of which will impact on the efﬁcient use of resources. Social production and ownership is the most usual way of addressing the problem with respect to new development. In the existing private rented sector rent controls tend to be the immediate response, but in the longer term the results of that control often generate a greater incentive for social provision and allocation (Quigley 1998).

The Economics of Social Housing

141

It can be reasonably argued that the most important reason for government intervention and provision is that housing is seen as a merit good – i.e. one where the value to society is higher than for individual members of society. This is based not just on the existence of externalities and distributional concerns, but because of interdependence of utilities – so it is often regarded as paternalism (Musgrave 1959). This argument applies most clearly at the bottom end of the system but can also apply more generally, if the accepted social welfare function includes signiﬁcantly higher housing aspirations (Hancock 1991). Thus there is no doubt that housing is subject to signiﬁcant market failures. These generate lower than optimal levels of housing investment and management as well as distortions in that provision, take inadequate account of the wider beneﬁts of good quality accommodation as well as the wider costs of bad housing and neighbourhoods, make it difﬁcult to address problems of risk and ﬁnancing to the individual, and do not allow available economies of scale and scope to be realised in order to minimise the costs of production.

Is the existence of market failure enough to support government provision? The analysis of market failure in housing generates an a priori case for government intervention – if not directly for government provision. Two important questions have to be addressed before the practical case is made:

•

is the impact of market failure in terms of higher costs, lower beneﬁts, under-investment and inappropriate allocation of housing resources large enough to matter in an anyway imperfect world; and

•

will administrative allocation actually overcome the problems of market failure or simply replace them with equally undesirable or even worse problems of administrative failure?

A great deal of work has been done over the last few years examining the costs of poor housing and the beneﬁts of increasing investment (Whitehead 1998; Bratt 2002). These have concentrated on two main aspects – the impact on lifetime opportunities and social exclusion, together with the associated costs to government, society and neighbourhoods of these failures; and on the positive role that housing can play in effective urban regeneration. There is growing evidence of a relationship between poor housing and poor health. Overcrowding, dampness, cold, poor air quality and housing-related

142

Housing Economics and Public Policy

stress all appear to have some relationship with poor health, the probability of accident and particularly the individual’s perceptions of problems (Kellet 1989; Anderson et al. 1993; Burrows et al. 1997; Carr-Hill 1997; Housing Studies, Special Issue 2000). However, except for those in temporary accommodation, the impact is not seen to be very great. The relationship between poor health and the experience of homelessness and living in temporary accommodation is much stronger (Burrows et al. 1997; Social Exclusion Unit 1998). However, the existence of a relationship is not enough to determine causality, especially as there is strong evidence to suggest that poverty and indeed inequality affect health and housing experiences (Wilkinson 1996). Similar analyses have been undertaken with respect to housing and education – where again the strongest evidence of the costs of poor housing relates to those who have experienced homelessness or have been placed in temporary accommodation (McBeath 1997; Stone 1997). The most important problems that have been identiﬁed tend to relate not so much to the quality of the home itself – except to some extent in terms of overcrowding – but more to the neighbourhood, and indeed the neighbours. Much the same applies to employment opportunities, where problems relate as much to peer group experience and accessibility to the appropriate labour market as to housing conditions themselves (Hopkins et al. 1997). An issue of considerable current relevance relates to the positive role that housing can play in effective urban regeneration. Early regeneration programmes in the UK concentrated almost entirely on housing investment, in part because subsidy arrangements could be readily put in place and because of its capacity to lever-in private ﬁnance. Latterly, regeneration programmes have become more widely based, taking in employment and training opportunities, programmes to reduce antisocial behaviour and improvements in social infrastructure. Even so, housing remains at the core of this more comprehensive approach because it is central to the interests of residents, and housing organisations tend to be relatively well-equipped to take leading roles in the partnerships necessary for success (Murie & Nevin 1997; Social Exclusion Unit 2001). Thus the evidence on the wider impact of poor quality housing and inadequate investment as well as the beneﬁts of housing-led regeneration points to signiﬁcant, if not overwhelming, market failure problems that generate costs both to the individuals involved and to the locality and society more widely. It is far less obvious that social investment in housing will necessarily solve these problems. It is clear that social housing programmes can generate

The Economics of Social Housing

143

large-scale investment both in new dwellings and in the existing stock – this is evidenced across Europe over the post-war period. It is less clear that the results are achieved at the lowest cost; that they are indeed additional, in that private housing development may be reduced, offsetting social investment; that the design and management of building are the most appropriate; that social housing meets people’s aspirations, in terms of the type of housing provided; and that the allocation process enables adequate choice and so maximises the value to the individual. In England it is also clear that the total amount of social provision is not adequate to ensure that deﬁned government objectives, in terms of a decent home for every household at a price within their means, can be achieved. In many areas there is a mismatch between the supply of housing and identiﬁed needs, resulting in continuing problems of homelessness and the use of temporary accommodation, even for those for whom the government takes direct responsibility. Moreover, social provision and allocation policies can, of themselves, generate signiﬁcant costs. In particular they tend to concentrate problems in speciﬁc localities resulting in major problems of social exclusion (Social Exclusion Unit 1998). In this context it has been argued that the problems are, to a signiﬁcant extent, the effect of administrative rather than market failure. On the other hand, although there are heavy concentrations in some social housing areas, the majority of such problems occur outside these areas, including in particular localities dominated by the private rented sector (Smith 1999). Equally, similar concentrations of problems are found when allocation is left to the market sector – as is obvious from US experience (Housing Studies, Special Issue 2002). The evidence on market failure reasons for social housing provision thus relates mainly to the potential external costs of bad housing and neighbourhood, and to the social value put on good quality accommodation. Other fundamental aspects of market failure, such as the extent of suboptimality in investment and allocation, have been subject to far less empirical scrutiny.

Housing as a means of redistribution The principles The strongest political case for the provision of social housing has never been in terms of efﬁciency, but rather in terms of a direct and effective means of redistribution and ensuring minimum housing standards. The main strand of the argument is straightforward: low-income households need assistance to purchase the necessities of life. How should this assistance be provided?

144

Housing Economics and Public Policy

One way is in the form of income subsidies, increasing capacity to pay but allowing freedom of choice about what is purchased. Success then depends on the efﬁciency of the market incentive system to generate the required supply response. Alternatively the subsidy can be given to providers, either private or social, directly to provide additional accommodation and to allocate it on the basis of agreed priorities. This then raises the additional question of how these priorities should be translated into practical decisions – in other words, who should receive the actual subsidy that is affected by both the allocation and pricing principles applied. A rather different distributional issue relates to the impact of disequilibrium on prices and allocation, especially where there are housing shortages, so that beneﬁts are being redistributed to the private landlord or landowner at the expense of consumers. The immediate problem can be addressed by rent controls in the private sector. However, these may adversely impact on the incentive to expand provision to bring provision back into equilibrium and imposes costs on those not able to access controlled sector housing. Equally it can be alleviated by social provision and rental subsidies for those in priority need. The choice might be argued to depend on the relative efﬁciency of the two sectors in responding to the incentive to provide as well as on the relative equity of the resultant allocation rules. The basic argument of principle is usually seen in terms of the relative beneﬁts of bricks and mortar as compared with income subsidies (Hills et al. 1990; Galster 1997; Yates & Whitehead 1998). The form of that subsidy may, in turn, impact on the amount of both public and private resources made available and on the relative capacity of the social and private sectors to ensure that provision is forthcoming. It might not be unreasonable to argue that one of the few lessons politicians appear to have taken from economists is that income subsidies are to be preferred to supply subsidies on the grounds that they provide higher utility to the individual because of the greater choice available. This is taken to apply without modiﬁcation to the case of housing. However, such an argument is over-simplistic. In principle regulation, taxation, subsidy and direct provision can all address market failures equally effectively. In reality appropriate methods of intervention depend both on the nature of the market and the practical attributes of intervention procedures. It is thus necessary at the same time to address the question of the relative efﬁciency of different means of intervention to offset market failures (Whitehead 1983). In particular income subsidisation takes no account of the relative capacity of the private and social sectors to provide the additional housing. It is also partial in that it only takes account of the housing market and not of other distortions such

The Economics of Social Housing

145

as the impact on the labour/leisure choice of recipients (especially where, as in the UK, the form of the income subsidy is particularly inappropriate). Equally, supply subsidies can be used to offset externality and other market failures, that are not directly addressed by choice-based demand subsidies. Supply subsidies are also the most obvious way of reﬂecting the extent to which housing is a merit good. Lastly, in a second-best world where there are distributional as well as efﬁciency objectives, it can be shown that transfers in kind of goods with low price elasticities of demand may generate fewer allocative distortions than income transfers (Bös 1983, 1985). Fundamentally, therefore, even though the question being addressed is that of distribution, the most appropriate way to achieve an agreed distributional objective comes down to the relative importance of different market failures on the ground. This can only be answered by empirical evidence. Within this more general question of income subsidies versus subsidies to provision as well as regulation in the face of shortage lies the role of prices and rents, which determine the extent of assistance provided. In a well-operating market environment prices reﬂect resource costs and therefore help to maximise the utility achieved from given housing resources. In practice, especially where there is disequilibrium, capacity to pay can come to dominate these efﬁciency beneﬁts and make intervention inevitable. In social housing the determination of rents and prices raises signiﬁcant problems about underlying objectives to help meet socially determined needs – e.g. by making larger housing cheaper so that those with families can pay for adequate accommodation – or to reﬂect the relative valuation or cost of the housing provided (Maclennan 1986c; Kleinman & Whitehead 1991). The problems of ensuring distributional as well as efﬁciency objectives are exacerbated when pricing below market levels generates shortages, so that those excluded have to pay far more for their housing in the market sector.

The practice It is hardly surprising that views about the relative beneﬁts of demand and supply subsidies in different countries depend signiﬁcantly on the government’s ideological stance (Yates & Whitehead 1998). In the USA there is general acceptance that demand-side subsidies are more efﬁcient – in part because there is little weight given to market failure, merit good and social welfare arguments – so that the only criteria for this assessment relate to choice and cost (Galster 1997). Much of the US analysis supporting demandside subsidies is based on an understanding of how people in receipt of subsidy might have the incentive to move between sub-markets and therefore

146

Housing Economics and Public Policy

provide incentives to suppliers to increase investment and remove poor quality units – a version of the ﬁltering theory. In Europe the basis for assessment tends to be wider. The analysis by Yates and Whitehead, for instance, argues for a more agnostic approach, based on an understanding of the factors preventing either form of subsidy effectively achieving their goals, more detailed assessment of the relative weights to be given to the different objectives and empirical evidence on their relative effectiveness. In practice in the UK and in most of the rest of Europe, although supply-side subsidies continue, increasingly they tend to be targeted. The extent of overall subsidy to housing has been much reduced and the importance of demandside subsidies has increased (Turner & Whitehead 2002). Thus, supply subsidies have been reduced to the point where in the Netherlands and Sweden, both countries with a strong commitment to housing as a social good, there is now no longer a net ﬁnancial allocation by government to supply, and in Britain such subsidies are limited to improvement of the social stock, together with a very limited new build programme. Arguably, this reﬂects the changing impact of market failure in growing economies. The large-scale development programmes have solved the problem of shortage, except in areas of sudden change (as in some parts of Germany in the 1990s). Increasing national income has generated higher average housing standards and reduced the impact of basic negative externalities, while the worst housing has been cleared or improved with government assistance. Individual demands, across much of the income distribution, are such that basic repairs and improvements are generally undertaken even if the level of investment is below the social optimum. As standards rise, the emphasis on housing as a merit good – as compared for instance with health, where improved technology keeps expanding the range of treatments and preventive medicines which society wishes to ensure are available to all – must inherently decline. Equally, national incomes have generally increased in absolute terms so that even if the actual distribution of income may have worsened so the need is seen to be for broader support to low-income households. As a result, only three main areas of intervention appear to remain important in terms of current government priorities in Europe:

•

ensuring access to adequate housing for those at the bottom of the income scale who are still excluded because of the failure of other programmes;

•

addressing the issue of neighbourhood, as the nature of housing services is redeﬁned to take account of local environment and services (where market failure arising from negative externalities tends to dominate), and

The Economics of Social Housing

147

offsetting problems of adjustment, for instance, addressing the housing consequences of sudden reductions in regional or sub-regional economic viability; and

•

maintaining and improving the existing stock, where problems of underinvestment arising from inadequate information and risk aversion are prevalent.

At the same time many European countries have been restructuring their social sectors and introducing greater private involvement, especially with respect to funding. In the Netherlands in particular the emphasis has been on ring-fencing the social rented sector and is now moving towards policies that aim to increase the proportion of owner occupation. In Sweden direct subsidies are being concentrated on particular areas of economic decline. In Britain the majority of council housing is being transferred to independent landlords, greater emphasis continues to be placed on shared ownership and other low-income homeownership schemes, and new social sector building is being directly linked to private sector developments (Turner & Whitehead 2002). In the USA the emphasis remains on providing direct assistance only to those at the bottom end of the income/access scale. Most new programmes are concentrated on increasing individual opportunity, reﬂecting the emphasis on choice. Equally, new programmes concentrate on offsetting the costs of immobility and discrimination (notably racial discrimination) that tend to exclude households from employment and housing markets. These policies particularly stress physical relocation into more mixed communities and access to owner occupation that is now seen as carrying with it many beneﬁts of access to credit, employment and a healthier lifestyle (Housing Studies, Special Issue 2002). Thus the general trends across Europe and North America are towards greater concentration of assistance, targeting low-income households and poor neighbourhoods. It is still the case however that the emphasis in Europe remains more on continued intervention in supply, especially in terms of regeneration, while in the USA assistance is concentrated on helping particular individuals to improve their position in the housing system. Both these approaches raise issues of the relative efﬁciency of private and social provision and thus the question of market versus administrative failure.

148

Housing Economics and Public Policy

The governance of housing The principles Throughout the 1970s the question of market failure, and therefore the case for government intervention, dominated the economic discussion of housing policy. It was not until the early 1980s that economists, led by political reality, started to apply the emerging analysis of comparative governance systems, and particularly that of administrative failure, to the question of housing provision. The theory of administrative failure is based on an examination of the nature of property rights. Well-deﬁned private property rights give the individual power to use resources to that individual’s best possible advantage. They will be used efﬁciently because that individual will wish to maximise independent individual utility. Private property rights also allow these resources to be transferred in such a way that expected future values are properly capitalised in the price. Thus resources will be put to their highest and best use and inefﬁciencies will be competed away. If this is the case then government policy should concentrate on liberalising markets and improving contractual frameworks. Social ownership carries with it problems of ill-deﬁned property rights and an incapacity to capitalise either efﬁciency or inefﬁciency. There is therefore no direct incentive to transfer resources to more efﬁcient providers. Inadequately deﬁned interaction between decision makers and those carrying out the policy generates inefﬁciencies of team production, because information is imperfect and decisions are based on bounded rationality, so that the outcome is not that desired (Hölmstrom 1979, 1982; Macho-Stadler & PérezCastrillo 1997). In particular there are inherent problems of agency that are not mediated by market pressures. Thus decision makers use resources to meet their own objectives rather than those of society or individual owners. These problems are exacerbated in situations of risk and uncertainty, perhaps particularly where that risk is political (Williamson 1979). Thus, local authorities and other social providers tend to have a paternalistic view of housing provision and allocation that does not necessarily reﬂect either individual or society objectives. Equally, the costs of provision, and especially management and maintenance, tend to be signiﬁcantly higher than would be the case under proﬁt maximisation. Allocation mechanisms will thus not normally reﬂect either opportunity cost or value. At the same time rent seeking behaviour tends further to misallocate resources (Posner 1972). These problems are seen to

The Economics of Social Housing

149

be particularly prevalent in housing because of the complex nature of the good and the different values that this complex mixture of housing attributes generates to those with different preferences. The more general analysis of transactions costs (Williamson 1975, 1986) stresses the extent to which internalising decisions within organisations can reduce the potential range of contractual failures – arguing for ﬁrms rather than markets and for integration where there are incomplete contractual arrangements. This analysis would tend to favour larger scale provision, internalisation of externalities and removing contractual arrangements – for instance through owner occupation rather than landlord/tenant arrangements. On the other hand problems of information and agency within the organisation and in particular the lack of external market pressures to allocate resources tells against social ownership arrangements. The core of the comparison between different forms of governance lies in three elements.

•

Clarity of objectives – if objectives are not transparent, and in particular, if there are multiple objectives, it is impossible to draw up effective contracts and it is extremely difﬁcult to specify incentives and to monitor outcomes effectively.

•

The appropriateness of incentive structures especially, where there are trade-offs to be made between risk and maximising values – administrative systems tend to have far less well-deﬁned incentive structures but to have a wider range of monitoring procedures available.

•

The framework by which the relevant activity – here the provision of social housing – is linked to the rest of the economy and therefore to overall resource allocation. In market systems this is done through asset capitalisation and realisation; in administrative systems it is achieved mainly through the regulatory framework.

In reality, any governance system is imperfect. It is therefore a matter of analysing the different attributes of the market, regulatory and administrative systems in the context of the particular product to determine the least-cost approach – as such, many of the most relevant factors come back to market and informational failures. Over the last few years there have been considerable developments in the analysis of ways to reduce the costs of governance failures of all types, notably in the context of ﬁnance markets but also in relation to non-proﬁt

150

Housing Economics and Public Policy

maximising organisations. These stress appropriate risk sharing procedures and also the use of debt ﬁnance and regulatory frameworks in order better to enforce efﬁciency incentives (Pollack 1985; Hansmann 1996; Allen & Gale 2000). All suggest that there is no single answer, but they also suggest that the form of governance does not have to be directly related to the extent to which social and private objectives differ (Bromwich et al. 2002). It is these developments that open up the opportunities for using the whole range of possible approaches – from public ownership and allocation through to private markets with income subsidies – to achieve the social goal of a decent home for every household at a price within their means.

The reality: administrative versus market failure It was not until after the Conservative government was elected in 1979 that the analysis of administrative failure began to emerge as central to the economic analysis of housing policy in the UK. At that time the sale of council houses through the Right to Buy was included in the Conservative Manifesto as the ﬁrst intimation of the priority to be given to more general privatisation (see the chapter in this volume by Williams). The reasons housing led this change in policy were mainly political and macroeconomic: ease of immediate cutbacks (85% of the net cuts in the ﬁrst two years of the Conservative government concentrated on housing); the potential for using the liberalised ﬁnance system both to extend owner occupation and to restructure social rented sector assets; and the fact that housing was not high on the political priority list. More fundamentally there was growing emphasis on individuality (as Mrs Thatcher famously said, there is no such thing as society) and the beneﬁts of choice, particularly in housing allocation where its private good nature appeared to have been lost in the actuality of social provision. Privatisation was therefore seen as improving productive efﬁciency, making resource allocation far more consumer oriented, and utilising the newly liberalised ﬁnancial system to introduce private incentives and constraints into production and later into the management and improvement of the existing stock (Whitehead 1984). This style of analysis lay at the heart of decisions to move to compulsory competitive tendering (Baker et al. 1992). It also lies at the heart of the private ﬁnance initiative that attempts to bring together the beneﬁts of lifetime private contracting and risk bearing with meeting well-deﬁned social objectives. Since 1997 these ideas have become more closely linked with a growing emphasis on models of partnership and consortia that bring together the required range of skills within an appropriate incentive framework. This is in line with the developing analysis of

The Economics of Social Housing

151

agency relationships and how to minimise the costs of mediating different objectives (Bromwich et al. 2002). However, there is still very little empirical analysis of the outcomes of changing governance structures. The early work by Duncan Maclennan on the relative efﬁciency of management in the local authority and housing association sectors remains one of the few detailed comparative studies (Centre for Housing Research 1989). In the main, evaluations have examined results on their own terms – for instance, what has been the outcome of introducing private ﬁnance into housing association provision (Chaplin et al. 1995). This in part reﬂects political reality – but it is a failure nonetheless. Moreover, even if administrative failures can be limited by better incentive systems and improved monitoring and enforcement, which would make it more worthwhile both to meet reasonable household demands and to use least-cost methods of production, there remains a fundamental problem. In part, because politically housing is not a priority, public funding will always remain constrained. As a result levels of provision are held below the optimum, and decisions about, for instance, rent increases and permission to borrow, that limit the capacity both to build the new units required or to maintain and improve the existing stock, are made not on the basis of optimal resource allocations but of political expediency.

Conclusions The analysis in this chapter suggests a practical rationale for observed trends across the industrialised world – but economic principles have had an impact, not only on immediate policy decisions but also on our general understanding of how and when governments should provide social housing. These principles also allow us to make a ﬁrst attempt at answering the questions about the nature and role of social housing posed at the beginning of this chapter: do we need a signiﬁcant social sector; should it be publicly owned; should rents be below market levels; and should administrative allocation continue? The answers are not straightforward in either economic or political terms, and in both contexts the appropriate way forward depends on the impact of past policies and the existing strength of organisational, regulatory and market frameworks. Both economic theory and evidence point to the need for signiﬁcant housing-speciﬁc subsidies to lower income housing to achieve social objectives. Current estimates for England would imply, for instance, that over one third of newly forming households will require some form of assistance (Holmans

152

Housing Economics and Public Policy

2001). Analysis suggests however that social objectives can be achieved without public, or indeed social, ownership. But, if we start from the current situation in the UK, were inappropriate constraints to be removed, both local authorities and housing associations would be likely to have a comparative advantage in providing, or ensuring provision into the longer term, for those in need of secure, subsidised accommodation. The case for rents below market levels is far less clear, especially in areas where there is a reasonable balance between households and dwellings. The ideal is rather to ensure provision is adequate, and income-related assistance is structured to give suitable incentives as well as reasonable access to housing. This can seem an impossible dream in current conditions, especially if economic activity remains so spatially unbalanced. Yet the economic case for greater consumer choice in an expanding economy where minimum standards of housing can be achieved from a reasonable proportion of basic incomes is strong. This also implies that, while the case for ensuring those with least bargaining power are helped to achieve acceptable housing is overwhelming, that for administrative allocation is much weaker. The most important implication of the analysis is that, in the twenty-ﬁrst century, there should be less emphasis on social housing per se and far more on managing markets, ensuring adequate investment and enabling different types of providers (public, non-proﬁt and private) to compete on cost and quality, rather than ideological grounds. A role for social housing will remain, but it should only survive if it both meets reasonable aspirations and provides good value for money. While changes in government policy in most countries can still best be described as knee-jerk responses to particular political pressures, there is evidence of rather slow movement towards a framework that attempts to bring together the better elements of private provision and allocation and link them with wider social objectives. In principle this could produce a social housing system using a range of instruments and relationships chosen on the basis of the relative costs of market and administrative failure. In practice, what is still the most likely outcome is that little will change, especially during periods when political pressures are low, but that political tensions, either macroeconomic or distributional, will dominate in times of crisis. Housing policy will then remain a cat’s cradle of policies addressing speciﬁc problems rather than a coherent whole – but at least economics has given us some tools by which we can assess outcomes and suggest improvements.

9 Neighbourhood Dynamics and Housing Markets George C. Galster Introduction The urban neighbourhood has been the subject of intense scrutiny by several academic disciplines since at least the 1920s. Economists came rather late to this ﬁeld, amassing what could be called a body of work only since the 1960s. Two volumes that Duncan Maclennan wrote and contributed to (Maclennan 1982; Grigsby et al. 1987) laid an important foundation for this emerging ﬁeld of economic study. Their contributions were to: (1) place the analysis of neighbourhood dynamics within the overarching context of metropolitanwide but segmented housing markets, and (2) demonstrate that standard neoclassical assumptions about supply, demand, and market operations required signiﬁcant supplementation in the case of housing and neighbourhoods.1 This essay builds upon these contributions, examining further the nature of neighbourhood, the source and nature of changes therein, and the consequences associated with such changes. I begin by forwarding an economics-based interpretation of neighbourhood, and consider the many idiosyncrasies associated with the concept. I go on to consider various aspects of the origins and nature of neighbourhood change,2 and argue that neighbourhood dynamics are rife with social inefﬁciencies. Lastly, I draw some inferences from the approach adopted for public policies related to neighbourhood regeneration.

Deﬁning neighbourhood I deﬁne neighbourhood as follows:

154

Housing Economics and Public Policy

Neighbourhood is the bundle of spatially based attributes associated with clusters of residences, sometimes in conjunction with other land uses. This deﬁnition owes its intellectual genesis to the work of Lancaster (1966), who originally formulated the notion of complex commodities as multidimensional bundles comprised of simpler (albeit sometimes abstract) goods. The spatially based attributes comprising the complex commodity called ‘neighbourhood’ consist of:

•

structural characteristics of residential and non-residential buildings: type, scale, materials, design, state of repair, density, landscaping, etc.;

•

infrastructural characteristics: roads, sidewalks, streetscaping, utility services, etc.;

•

demographic characteristics of the resident population: age distribution, family composition, racial, ethnic, and religion types, etc.;

•

class status characteristics of the resident population: income, occupation and education composition;3

•

tax/public service package characteristics: the quality of public schools, public administration, parks and recreation, etc. in relation to the local taxes assessed;

•

environmental characteristics: degree of land, air, water, and noise pollution, topographical features, views, etc.;

•

proximity characteristics: access to major destinations of employment, entertainment, shopping, etc., as inﬂuenced by both distance and transportation infrastructure;4

•

political characteristics: the degree to which local political networks are mobilised, residents exert inﬂuence in local affairs through spatially rooted channels or elected representatives;5

•

social-interactive characteristics: local friend and kin networks, degree of inter-household familiarity, type and quality of inter-personal associations, residents’ perceived commonality, participation in locally based voluntary associations, strength of socialisation and social control forces and local social institutions, etc.;6 and

Neighbourhood Dynamics and Housing Markets

•

155

sentimental characteristics: residents’ sense of identiﬁcation with place, historical signiﬁcance of buildings or district, etc.

The unifying feature of these attributes constituting the bundle called neighbourhood is that they are spatially based. The characteristics of any attribute can be observed and measured only after a particular location has been speciﬁed. This is not to say that neighbourhoods are homogeneous on any attribute, merely that a distribution or proﬁle can only be ascertained once a space has been demarcated. Moreover, to say that attributes are spatially based does not mean that they are intrinsically coupled with the geography; some are (infrastructure, topography, buildings), whereas others are associated with individuals who lend their collective attribute to the space purely through aggregation (race, income, life-cycle stage). I emphasise that while most of the attributes above are usually present to some extent in all neighbourhoods, the quantity and composition of constituent attributes typically vary dramatically across neighbourhoods within a single metropolitan area. This implies that, depending on the attribute package they embody, neighbourhoods can be distinctly categorised by type and/or by quality. This is, of course, a tenet of social area analysis (Greer 1962; Hunter 1974). However, unlike that school of thought, I extend the dimensions over which neighbourhoods can be classiﬁed beyond the demographic- and status-related. The extension is necessary if one is to understand neighbourhood change, for key decision makers evaluate more than merely the demographic and status attributes of a space before investing in it. Commodities are consumed, of course, but in the case of neighbourhood many of the attributes take on a quasi-public good character, with nonexcludable, non-rival consumption. Four distinct types of users potentially reap beneﬁts from the consumption of neighbourhood: households, businesses, property owners, and local government.7 Households consume neighbourhood through the act of occupying a residential unit and using the surrounding private and public spaces, thereby gaining some degree of satisfaction or quality of residential life. Businesses consume neighbourhood through the act of occupying a non-residential structure (store, ofﬁce, factory), thereby gaining a ﬂow of net revenues or proﬁts associated with that venue. Property owners consume neighbourhood by extracting rents and/ or capital gains from the land and buildings owned in that location. Local governments consume neighbourhood by extracting tax revenues, typically from owners, based on the assessed values of residential and non-residential properties.

156

Housing Economics and Public Policy

Idiosyncrasies of neighbourhood Here I discuss four aspects of spatially based attributes and the idiosyncrasies that result for neighbourhood dynamics. These aspects are: cross-attribute variation in durability, cross-attribute variation in ability to be priced, relativistic evaluations of attributes by consumers, and consumption impacts on attributes. The spatially based attributes comprising neighbourhood vary in their durability. Some, like certain topographical features, are permanent. Sewer infrastructure and buildings typically last generations. Others, such as tax/public service packages and demographic and status proﬁles of an area, can change over a year. The area’s social interrelationships and public behaviours can sometimes alter even more rapidly. The implication of this observation is as follows. Although some of the key features that deﬁne a desirable neighbourhood from the perspective of its many consumers can be counted on to remain constant (and therefore predictable) for extended periods, others cannot. This means that consumers’ predictions about future changes in these less-durable features, coupled with their aversion to risk and their preferences for these various, non-durable attributes, will play a major role in determining decisions about mobility, ﬁnancial investments, and psychological investments in neighbourhoods over the long term.8 The spatially based attributes comprising neighbourhood vary in their ability to be priced by market mechanisms. In order for potential consumers to make bid offers for a commodity they must have some modicum of information about the quantity and quality of that commodity and what likely beneﬁt they would receive from its consumption. Real estate markets have been shown to meet this criterion for a large number of spatially based attributes. Indeed, this is the foundation of over three decades of empirical work estimating ‘hedonic indices’.9 These studies have shown that attributes like structural size and quality, accessibility, tax/public service packages, demographic and status composition of residents, and pollution can be priced and, in turn, be converted into a measure of household consumers’ willingness to pay. However, most social interactive dimensions of neighbourhood cannot be priced well because they are hard ex ante for prospective bidders to assess. The idiosyncratic and personalised nature of

Neighbourhood Dynamics and Housing Markets

157

neighbourhood social interactions means that prospective in-movers will only be able to ascertain how they will ‘ﬁt in’ after an extended period of residence. One implication is that long-term residents may have considerably different market evaluations (‘reservation prices’) for their neighbourhood than prospective residents or investors because the former have capitalised (positively or negatively) their assessments of the social interactive dimension. Thus, the former may be highly resistant to external market forces when they assess a positive social environment, and may be more easily out-bid and eventually supplanted by new owners and residents when they assess a negative one. Another implication is that neighbourhoods are particularly prone to forms of insider dealing, with privileged information communicated to preferred buyers and in-movers by current residents, owners, and their market intermediaries. Even if the market can price attributes comprising neighbourhood, however, their price will typically be based on a comparison of attributes in competing neighbourhoods, not on the intrinsic characteristics of the attribute set.10 Perhaps the most obvious example is the status dimension. The absolute income levels of households in a particular neighbourhood may rise, but if they are rising at least as quickly in all other neighbourhoods in the metropolitan area there likely will be no change in consumers’ evaluations of that neighbourhood’s status attribute. Analogous arguments can be made regarding other attributes, such as proximity, school quality, and public safety. The upshot is that, when new neighbourhoods are created through large-scale construction or rehabilitation projects, they can change the relative attractiveness of existing neighbourhoods. And because relative evaluations will alter ﬂows of resources across space, absolute changes in the existing neighbourhoods will follow (Galster 1987; Grigsby et al. 1987). The spatially based attributes comprising neighbourhood can change by the very act of consuming them. This can occur directly and indirectly. With regard to direct change, as households consume neighbourhood by occupying residences in it, they may simultaneously alter the demographic and/or socio-economic status proﬁle of the neighbourhood if the in-moving households differ systematically from longer-term residents. Analogously, a different type of ownership of homes or stores may emerge if the consumption changes to absentee-owned instead of owner-occupier, for

158

Housing Economics and Public Policy

example. Indirectly, changes in the occupancy and/or ownership proﬁles of a neighbourhood not only changes tautologically its current attributes, but may trigger longer-term changes in a wider variety of attributes. This can occur if the occupancy or ownership changes yield different decisions by current or prospective consumers of the neighbourhood that affect the ﬂows of resources into that space, a topic I discuss in more depth later. Put differently, the monetary and psychic beneﬁts and costs associated with the consumption of a neighbourhood largely originate externally to the consumer in question (Maclennan 1982). Sufﬁce it to note brieﬂy here the corollary of the foregoing point: attributes of neighbourhood are mutually causal over time. Changes in one attribute may change decisions by one or more type of consumer, which lead, in turn, to changes in other attributes, and so on (Galster 1987; Grigsby et al. 1987; Temkin & Rohe 1996).

How neighbourhoods come to be Although it is tempting to conceive of neighbourhood as a commodity with ﬁxed, clearly deﬁned characteristics, it is more appropriately viewed in a more dynamic perspective. The attributes comprising neighbourhood at any moment are, in fact, the result of past and (typically) current ﬂows of households and resources – ﬁnancial, social-psychological, and time – into and out of the space in question (Galster 1987). Certainly, when a subdivision or estate of homes is newly constructed one might say that a neighbourhood has come into being, though without household occupants it is not yet a fully formed neighbourhood. From that moment on, what attributes that place will possess – what that neighbourhood will be – will be shaped by the decisions of current and prospective consumers. Thus, in a fundamental way, the consumers of neighbourhood can be considered the producers of neighbourhood as well. Households consume a neighbourhood by choosing to occupy it, thereby producing an attribute to that location related to that household’s demographic characteristics, status, civil behaviours, participation in local voluntary associations and social networks, and so forth. Property owners consume a neighbourhood by buying land and/or buildings in it; they subsequently produce the neighbourhood’s attributes through their decisions regarding property construction, upkeep, rehabilitation, or abandonment. Business people consume a neighbourhood by operating ﬁrms there, thereby producing attributes related to structure types, land use, pollution and accessibility. Local governments consume

Neighbourhood Dynamics and Housing Markets

159

neighbourhood by extracting property tax revenue and, in turn, produce attributes associated with public services and infrastructure. The list of producers of importance to any neighbourhood is expansive and diverse. In the ﬁrst rank are the aforementioned consumers: households, property owners, business people, and local governments. The list of producers also includes, in a secondary but nevertheless important way, those in the real estate brokerage, insurance, and mortgage ﬁnance sectors. Within these broad categories of producers further dimensions of variation are manifested. These groups include both those who currently reside, own property, and/or earn income or tax revenues in a given neighbourhood, and those who do not but may under certain circumstances. They include those who perceive a vested ﬁnancial or social-psychological interest in the area and those who do not. They include those who make decisions using the cold calculus of proﬁt maximisation, those who consider sentiment and personal satisfaction, and still others who are motivated by political pressures. The mobility, purchasing, and resource allocation decisions related to neighbourhoods (what I will refer to hereafter as ‘investing’) are inherently fraught with an unusual amount of uncertainty emanating from both within and without. First, the aforementioned dual functions of consumer and producer mean that the decisions by one will affect directly and indirectly the investment outcomes of all in the neighbourhood. Although the neighbourhood as a whole affects the well-being of each consumer/producer, what happens to that neighbourhood is a function of changes in numerous constituent attributes associated with it, each of which has an uncertain future because there is a large number of consumer/producers of different types and motivations controlling it. Second, the ﬂows of resources across all neighbourhoods in a metropolitan area will be inﬂuenced by uncertain metro-wide factors related to the regional economy, technological innovation, population and immigration, state and federal government policy, and vagaries of nature (Grigsby et al. 1987; Temkin & Rohe 1996), as will be explained below. This high uncertainty translates into substantial long-term risk because, once made, investments of resources in neighbourhoods are not easily reversible. Many sorts of potential investments have substantial out-of-pocket and psychological transaction costs, costs which consumer/producers are loath to incur on a frequent basis. Other sorts of potential investments, especially structures and infrastructures, have long projected life spans and are spatially ﬁxed. The high uncertainty–high risk nature of neighbourhood investments holds important implications for the characteristics associated with neighbourhood change.

160

Housing Economics and Public Policy

Changes in neighbourhood Above I argued that neighbourhoods would change (i.e., their attributes would be altered) based on the decisions made by the welter of consumer/ producers that inﬂuence the ongoing ﬂow of resources to a neighbourhood. These decisions are based heavily on relativistic, inter-neighbourhood comparisons and future expectations embedded within a risk-laden, highly interactive, multi-actor context. In this section I will examine further the nature of neighbourhood dynamics these conditions give rise to. I ﬁnd it helpful in understanding why and how an individual neighbourhood changes to think of two overlapping processes at work. One is a macro process that reverberates through the entire system of neighbourhoods in a metropolis, which can best be analysed through a neoclassical, segmented market framework. The other is a series of micro processes describing the nature of adjustments within a neighbourhood and between neighbourhoods within a common market segment. The micro processes can best be modelled with ‘complexity models’ of non-linear and discontinuous changes. I begin below describing the macro process and argue that, fundamentally, forces originating outside of neighbourhoods initiate neighbourhood change. I then discuss micro-processes, and show how they result in non-linear, threshold adjustments in individual neighbourhoods. Lastly, I explain why changes in neighbourhood are socially inefﬁcient.

The macro process of neighbourhood change Consider a quality-segmented model of the metropolitan housing market.11 It presumes a predominantly private housing market for rental and sales units, wherein suppliers and households are motivated importantly (but not solely) by longer-term rates of return and satisfaction, respectively. Begin by classifying the housing stock in a metropolitan area into ‘quality sub-markets’, sets of dwelling units that households perceive as closely substitutable, considering all the various attributes of the housing bundle, including spatially based attributes comprising its neighbourhood (see the chapter by Tu in this book for a detailed consideration of the concept of housing submarkets).12 Sub-markets can be arrayed from ‘lowest’ to ‘highest’ quality levels, with increasingly dissimilar comparisons of quality levels representing decreasing degrees of substitutability.13 Each sub-market can be modelled as having its own supply and demand functions. It is beyond the scope of this chapter to consider the determinants of these functions; it will sufﬁce here

Neighbourhood Dynamics and Housing Markets

161

to emphasise only the mechanism of inter-sub-market repercussion, which is crucial for understanding the systemic process of neighbourhood change. Supply into one sub-market (through new construction and/or net conversion of existing dwellings) will be inﬂuenced, among other things, by the relative (risk-adjusted) rate of return that developers and owners can expect to reap in this sub-market compared to others. Demand by households in any one sub-market will be inﬂuenced, among other things, by the relative market valuations (sales prices or equivalent capitalised rents) in closesubstitute sub-markets. Shocks to either supply or demand in any one submarket are transmitted sequentially throughout the sub-markets array by housing owners/developers altering their supply decisions in response to a new sub-market’s relative pattern of expected rates of return, and by households altering their occupancy decisions in response to new relative market valuations across substitute sub-markets. This model of segmented metropolitan housing markets can be usefully applied to elucidating the fundamental macro process behind neighbourhood dynamics (Rothenberg et al. 1991). Metropolitan housing sub-markets and neighbourhoods are not synonymous, but their relationship is straightforward. A sub-market is more than a neighbourhood or contiguous collection of them. It is comprised of all closely substitutable dwelling units, regardless of their location or propinquity. However, most neighbourhoods in the US and many other nations consist primarily of residences classiﬁed (by households, owners, and developers) in the same quality sub-market, for three reasons (Vandell 1995). First, economies of scale in construction lead developers to build homes in a subdivision or units in large rental estates that typically have similar physical characteristics. Second, the American (and, perhaps, other societies’) willingness to pay premiums for class homogeneity, often backed up with various exclusionary land use regulations, limits diversity of housing types within small geographic areas (Grigsby et al. 1987). Third, because spatially based attributes contribute to a housing unit’s quality and, hence, sub-market, units in close proximity will share many common attributes and thus tend to be classiﬁed in the same quality sub-market tautologically. The foregoing suggests ﬁrst that macro forces strongly affecting a particular housing sub-market will similarly affect the neighbourhoods where such a classiﬁcation of dwelling is located; the greater the concentration of the given sub-market type in a neighbourhood, the greater the spatial impact there. Second, it suggests that macro forces originally impacting anywhere (either in terms of quality sub-market or geographic location) in the metropolitan area will eventually have some (albeit minimal, perhaps) impact

162

Housing Economics and Public Policy

everywhere, as the shock is transmitted in progressively damped severity across sub-markets (and the neighbourhoods where they are predominantly represented) of increasingly dissimilar substitutability. As illustration, consider the following scenario. Developers speculate by building neighbourhoods comprising a number of high-quality sub-market homes on undeveloped land at the periphery of the given urbanised area (Galster & Rothenberg 1991). Should this increase in supply exceed the increase in demand for the high-quality sub-market (say, due to growth of high-income households in the metropolitan area), there will be a net decline in the market valuations and rate of return associated with such dwellings, at least for some period. Some households who previously would have chosen not to occupy the high-quality sub-market now would do so, as affordability has risen from their perspective.14 Concomitantly, these declines in actual (and expected) rates of return would provide incentives for owners of pre-existing (especially rental) dwellings in the high-quality sub-market to downgrade the quality of their units to take advantage of comparatively superior rates of return in the somewhat lesser-quality sub-market(s). They typically accomplish this by passive under-maintenance: investing insufﬁcient upkeep to maintain the dwelling in its original sub-market. These adjustments by households and suppliers jointly work toward re-establishing the initial rates of return in the high-quality sub-market but simultaneously reduce those in the lower-quality one(s). There, demand has now fallen (from some erstwhile occupants choosing instead a superior quality sub-market) and supply has risen (from some owners downgrading from higher quality submarkets), thereby driving down market valuations. An adjustment process on both supply and demand sides of the market ensues analogous to that above, but disequilibrating forces are transmitted still farther down the sub-market quality array. At every step, the magnitude of inter-sub-market repercussion will depend on a host of elasticity-inﬂuencing factors implicit in the aggregate proﬁles of households (e.g. degree of positive social attachments to their neighbourhood, preferences for structural compared to neighbourhood components, risk aversion) and housing suppliers (e.g. cost of converting structures, expectations about duration of altered rates of returns, risk aversion). As the system as a whole adjusts to eliminate large inter-sub-market differentials in rates of return, the model predicts a series of changes in demographic and physical attributes of neighbourhoods constituting sub-markets. There is a tendency in every sub-market where demand and/or supply has been altered for the least competitive neighbourhoods to evince: (1) an in-migration of households of somewhat lower means than the typical residents who left (called ‘income succession’), and (2) a decline in the physical quality of the

Neighbourhood Dynamics and Housing Markets

163

dwellings; in the extreme, dilapidation and even abandonment may occur in the neighbourhoods where the lowest-quality sub-market predominates. The new construction of high-quality dwellings in excess of household demand for such rendered an array of somewhat lower-quality neighbourhoods relatively less attractive to households, and less expensive. This generated altered ﬂows of resources (occupancy patterns by households, ﬁnancial resources by owners) that ultimately changed absolutely the population and housing attributes of these neighbourhoods.15

The dominance of external forces of change What I wish to stress about the scenario described above is that forces originating outside of them triggered alterations to existing neighbourhoods.16 Although this particular scenario involved speculative new construction as the initial shock, equally powerful macro process dynamics can be produced by such things as metropolitan immigration or emigration, technological changes, new transportation infrastructure, alterations in taxes, utility costs, a wide range of public policies, and shifts in the household income distribution. These forces external to a particular neighbourhood may well lead those controlling resources ﬂowing into it to change their decisions, based on the connections described by the quality sub-market array of the metropolitan housing market. Thus, the most fundamental sorts of neighbourhood change are externally induced.17 This is not to say that, once initiated, change cannot become cumulative or self-fulﬁlling.

Micro processes of neighbourhood change Once triggered by macro process, the dynamic of neighbourhood adjustment is often non-linear, even discontinuous: what I call a threshold effect.18 There are four distinct, not mutually exclusive, mechanisms suggested by theory through which thresholds may be produced: collective socialisation, contagion, gaming, and preference models.19 The ﬁrst two rely upon collective actions and social intercourse to create thresholds; the other two involve more atomistic attitudes and behaviours. One can analyse: behaviour of households to move out through collective socialisation, gaming, and preference models; behaviour of households to move in through gaming models; and behaviour of residents, owners, and business people who remain in the neighbourhood through collective socialisation, gaming and contagion models.

164

Housing Economics and Public Policy

Collective socialisation theories focus on the role that social groups exert on shaping an individual’s attitudes, values and behaviours (e.g. Levine 1971; Weber 1978). Such an effect can occur to the degree that: (1) the individual comes in social contact with the group, and (2) the group can exert more powerful threats or inducement to conform to its positions than competing groups. These two preconditions may involve the existence of a threshold. Given the importance of interpersonal contact in enforcing conformity, if the individuals constituting the group in question were scattered over urban space, they would be less likely to be able either to convey their positions effectively to others with whom they might come in contact or to exert much pressure to conform. It is only when a group reaches some critical mass of density or power over a predeﬁned area that it is likely to become effective in shaping the behaviours of others. Past this threshold, as more members are recruited, the group’s power to sanction non-conformists probably grows nonlinearly. This is especially likely when the position of the group becomes so dominant as to become normative in the area.20 The basic tenet of contagion models is that if decision makers live in a community where some of their neighbours exhibit non-conforming behaviours, they will be more likely to adopt these behaviours themselves. In this way, social problems are believed to be contagious, spread through peer inﬂuence. Crane (1991) proposes a formal contagion model to explain the incidence and spread of social problems. He contends that the key implication of the contagion model is that there may be critical levels of incidence of social problems in neighbourhoods. He states that if ‘the incidence of problems stays below a critical point, the frequency or prevalence of the problem tends to gravitate toward some relatively low-level equilibrium. But if the incidence surpasses a critical point, the process will spread explosively. In other words, an epidemic may occur, raising the incidence to an equilibrium at a much higher level’ (p. 1227). Gaming models assume that, in many decisional situations involving neighbourhoods, the costs and beneﬁts of alternative courses of action are uncertain, depending on how many other actors choose various alternatives. The individual’s expected payoff of an alternative varies, however, depending on the number or proportion of others who make a decision before the given actor does. Thus, the concept of a threshold amount of observed prior action is central in this type of model. The well-known prisoners’ dilemma is the simplest form of gaming model (Schelling 1978), but more sophisticated variants have been developed and applied to neighbourhood change processes (Granovetter 1978; Granovetter & Soong 1986). As illustration, consider the situation of a dilapidated neighbourhood for which the market

Neighbourhood Dynamics and Housing Markets

165

is signalling potential gains in property values were its owners to improve their properties as a group. But individual owners may believe that they will not earn back the value of their marginal investment if they were to upgrade but no others followed suit. A conservative gaming strategy of behaving to minimise maximum prospective loss, regardless of what others may do, will lead many owners to refrain from upgrading ﬁrst. Only if a threshold proportion of owners were to upgrade would these sceptics be convinced to upgrade (Taub, Taylor & Dunham 1984). Preference models claim that actors in a residential environment will respond if the aggregate behaviour of others (or an exogenous event) raises an undesirable neighbourhood attribute above the level they ﬁnd tolerable. An endogenous process can be triggered once the attribute reaches the critical threshold. The trigger occurs because actors in a neighbourhood are assumed to have different tolerance levels, with the least tolerant responding ﬁrst. If additional change in the neighbourhood attribute results from the course of action taken in response to the initial event by those with the lowest tolerance level, the new level of the neighbourhood attribute may now be above the tolerance level of some of the less tolerant remaining actors. The process may continue with new rounds of attribute change and actor adjustment until the process is completed. At the extreme, the process may end when all the original actors in a neighbourhood have responded. The theoretical development of preference models has focused on changes in a neighbourhood’s racial composition, though extensions to preferences for other sorts of neighbourhood attributes are straightforward.21 Uncertainty over the origin of thresholds should not obscure the point that the dynamic of neighbourhood adjustment initiated by macro process is not necessarily swift, smooth, or continuous. Micro processes introduce considerable complexity and uncertainty about adjustments. In some cases, neighbourhoods will be extremely slow to adjust; in other cases they will adjust spasmodically if they began near their threshold point. Macro process may induce a few households to change neighbourhoods, but if the movers’ race or class matters signiﬁcantly to households in either origin or destination neighbourhoods the adjustments associated with micro sorting processes may go far beyond what would have been predicted at the macro level.

Neighbourhood change and social efﬁciency The foregoing discussion implies that changes in the ﬂows of resources into neighbourhoods are not likely to produce socially efﬁcient outcomes. At

166

Housing Economics and Public Policy

least four reasons come to bear: externalities, gaming, expectations, and inadequate pricing of attributes. Because the act of consuming neighbourhood can change its attributes directly and indirectly, and such changes affect the decisional calculus of other consumer/producers, the act can be thought of as generating externalities. The choice of a minority to move into an all-white occupied neighbourhood imposes externalities on the bigots there. The choice of a property owner to repair the façade provides external beneﬁts to neighbours. Because such external costs and beneﬁts do not accrue to the decision maker, a sub-optimal amount of the activity is chosen. The earlier discussion of gaming is also once again relevant here. With lack of certainty about the decisions of other consumers/producers in the neighbourhood, yet dependency of one’s payoffs from alternative choices dependent upon such, autonomous decision makers are likely to adopt strategies that do not produce the greatest good for the collective. The unwillingness to renovate dilapidated buildings in an area until other investors do so ﬁrst is a classic example. Expectations are, of course, imperfect and prone to major errors. But this in itself does not imply a systematic bias toward inefﬁcient choices. Rather, expectations about the future may prove to be so ‘certain’ in the view of the decision maker that the resulting choice encourages the expectation to transpire. This is the famous ‘self-fulﬁlling prophecy’ phenomenon. An illustration is panic selling of homes. Because of some anticipated neighbourhood change, several homeowners become convinced that property values will fall rapidly. They therefore try to sell their homes quickly, offering a discount. But the rash of ‘For Sale’ signs and the rumours that these homes are selling cheaply convinces other owners in the neighbourhood that, indeed, values are on the way down. As they join in the attempt to unload their properties, panic ensues and prices do, as some prophesied, drop precipitously. The sorts of prices produced by these self-fulﬁlling prophecy dynamics are unlikely to allocate resources efﬁciently. Instead of accurately capitalising the underlying quality (and replacement cost of the dwellings) in the neighbourhood, these artiﬁcially-deﬂated prices encourage purchase by owners with less personal ﬁnancial means. These owners are likely to invest less in home maintenance and repair activities than their higher-income forebears (Galster 1987), thereby shortening the useful lifetimes of these valuable assets. Lastly, I explained above how certain attributes of the neighbourhood, especially those associated with the sentimental and social-interactive dimensions, could not be evaluated well by potential consumers/producers not

Neighbourhood Dynamics and Housing Markets

167

yet located in a given neighbourhood, compared with those located there for some time. This divergence in information creates an agency problem in which inefﬁcient transactions are likely to occur. Owners attempting to sell or rent their properties will have a bias toward disguising any undesirable attributes – and not discounting the price appropriately – of the neighbourhood that may not be known to prospective buyers or renters. This means that the latter group will make inefﬁcient choices: they would have chosen a neighbourhood offering a superior quality of life for the money had they but been fully informed.

Implications for neighbourhood regeneration policy Space permits only brief reference to some salient neighbourhood policy implications from the foregoing.22 First, thresholds imply a need for careful targeting of policy. Interventions concentrated on neighbourhoods poised on the brink of incipient, catastrophic decline represent a wiser allocation than spreading scarce public resources more evenly across an array of neighbourhoods. As illustration, my reading of the US empirical work on the effects of concentrated poverty suggests that a variety of social problems escalate when the neighbourhood poverty rate exceeds 20% (Galster, forthcoming). Second, decision-makers’ relativistic comparisons among neighbourhoods raise the spectre of zero-sum interventions. In a stagnant metropolis with little household or income growth, a massive public initiative to raise dramatically the physical quality of one depressed neighbourhood will erode the fortunes of all neighbourhoods over which the regenerated community leapfrogs. Inasmuch as the relative quality of such leapfrogged neighbourhoods have declined, investors and households with means will switch their resources from them into the regenerated area, leading to absolute decline of the former set. Several caveats are however in order. The consequences of regeneration of low-quality neighbourhoods will more likely have positivesum consequences to the degree that there is: (1) an increase (decrease) in the number of households with rising (falling) real incomes in the metropolitan area; (2) a substantial share of subsidised, means-tested housing provided in the regenerated area; (3) social housing providing the dominant form of supply in the leapfrogged neighbourhoods; and (4) a limited number of neighbourhoods leapfrogged because many are regenerated concomitantly.

Summary and conclusion In this chapter I have attempted to build upon the insights on the economics

168

Housing Economics and Public Policy

of neighbourhood provided two decades ago by Duncan Maclennan. I proposed that the urban neighbourhood could be usefully deﬁned as the bundle of spatially based attributes associated with clusters of residences, sometimes in conjunction with other land uses. This bundle of attributes is multidimensional, consisting of everything from structures and topography to demography, public services, and social interactions. The stock of attributes comprising neighbourhood at any moment is the result of past and current ﬂows of households and resources – ﬁnancial, social-psychological, and time – into and out of the space in question. Four key users make decisions affecting these ﬂows: households, businesses, property owners, and local government; they thus can be viewed as producers of neighbourhood. However, these same actors play dual roles because they potentially reap beneﬁts from the consumption of this complex commodity called neighbourhood. To understand the factors and processes that inﬂuence these production and consumption decisions is to uncover the roots of neighbourhood change. Multi-dimensional commodities are common, but the neighbourhood offers a variety of signiﬁcant idiosyncrasies that suggest the following insights into these decisions.

•

Different neighbourhood attributes vary in their durability. Investors therefore must take a long-term strategy with durable features, with concomitant reliance upon imperfect and socially inﬂuenced expectations. Self-fulﬁlling prophecies and gaming strategies that result yield inefﬁcient outcomes.

•

Social-interactive and sentimental attributes of neighbourhood are not well-priced because of information asymmetries. The resultant agency problem leads to inefﬁcient outcomes.

•

Consumers evaluate neighbourhood attributes relatively. This provides the vehicle by which changes elsewhere in the metropolitan area can lead to changes in the given neighbourhood. This implies that the prime origins of change in a particular neighbourhood are located outside that neighbourhood.

•

The act of households or property owners consuming neighbourhood attributes typically changes the attributes directly and indirectly. This means that attributes of neighbourhood are mutually causal over time. Because this process generates externalities imposed on other

Neighbourhood Dynamics and Housing Markets

169

consumers/producers operating in the neighbourhood, inefﬁcient outcomes are manifested. The foregoing suggests an unmistakable case of market failure. For a variety of reasons inherently associated with the concept of neighbourhood, changes in ﬂows of households and resources across space will produce socially inefﬁcient outcomes. There is thus a prima facie case for some sort of collective intervention, whether it comes from informal social processes, non-proﬁt, community-based organisations, or the governmental sector. Informal social processes might take the form of sanctions and rewards meted out by neighbours that are designed to enforce compliance with collective norms regarding civil behaviour and building upkeep. Community-based organisations might politically organise, establish neighbourhood bonds of mutual solidarity, or foment a positive public image of the neighbourhood. Unfortunately, the foregoing two options operate at too small a spatial scale to affect the macro process producing neighbourhood change. Governments might offer ﬁnancial incentives, regulations, and investments of infrastructure and public services, and target them to neighbourhoods at crucial threshold points. In concert, these actions can help alter perceptions of key neighbourhood investors, provide compensatory resource ﬂows, minimise destructive gaming behaviours, internalise externalities, and moderate expectations, thereby defusing self-fulﬁlling prophecies. On the other hand, massive regeneration efforts may run afoul of zero-sum neighbourhood effects unless particular criteria are met.

Acknowledgments I am grateful for the insightful suggestions provided on an earlier draft by Glen Bramley, Kenneth Gibb, Ade Kearns, Tony O’Sullivan, and Gavin Wood, that substantially improved this work. Some of the material in this chapter originally appeared in Urban Studies, whose permission to reproduce is acknowledged.

Notes (1)

Similar perspectives were also developed in roughly contemporaneous works by Galster (1987), and Rothenberg et al. (1991), that expanded the analysis theoretically and empirically.

170 (2)

(3) (4) (5) (6) (7)

(8( (9) (10)

(11)

(12)

(13)

(14) (15)

(16) (17)

(18)

Housing Economics and Public Policy For comprehensive reviews of the social scientiﬁc literature on neighbourhood, see Hunter (1979), Schwirian (1983), Hallman (1984), and Temkin & Rohe (1996). Von Boventer (1978) and Maclennan (1982) have been especially incisive in analysing the implications of this component. This was the salient factor determining household location choice in the earliest economists’ models (Maclennan 1982). For more on this dimension, see Hunter (1979), Schoenberg (1980), and Temkin & Rohe (1996). For more on this dimension, see Warren (1975), Schoenberg (1980), Fischer (1982), and Warren & Warren (1977). Visitors may also consume neighbourhoods in which they do not reside by the act of working, shopping, or seeking entertainment there. For simplicity I omit them as key consumers when analysing the main determinants of neighbourhood change. This point has also been stressed by Leven et al. (1976); Maclennan (1982). For critical reviews of theory and evidence on hedonic indices, see Maclennan (1977b; 1982); Rothenberg et al. (1991). Although some attributes may have associated with them an absolute minimum threshold value, below which no price will ever be bid, such as the case of air quality. This model was originally developed by Rothenberg et al. (1991). In housing markets dominated by state production and allocation mechanisms, the model has more circumscribed relevance as a tool for understanding neighbourhood change. Maclennan (1982) suggested that ‘sub-markets be adopted as a working hypothesis’, based on the durability, multi-dimensional complexity, and spatial ﬁxity of housing. He did not specify that quality be the metric for segmentation, however. See Malpezzi (this volume) for how hedonic indices can be employed in market segmentation studies. This proposition is true for both demanders and suppliers; owners of existing units ﬁnd it increasingly costly to modify their dwellings as the quality change implied grows. The precise response would be governed by the cross-price elasticity of demand between the sub-markets in question (Rothenberg et al. 1991). The astute reader will recognise this scenario as producing what often has been labelled ‘ﬁltering’. For more complete analyses, see Maclennan (1982) and Galster & Rothenberg (1991). A similar position was articulated in Maclennan (1982), Grigsby et al. (1987), Galster (1987) and Temkin & Rohe (1996). This stands in contrast to some who argue that the internal institutional and social structures of the neighbourhood are central to understanding its trajectory (Schoenberg 1980). This was also recognised by Maclennan (1982).

Neighbourhood Dynamics and Housing Markets (19)

(20)

(21)

(22)

171

For a review, see Quercia & Galster (2000); for new evidence, see Galster et al. (2000). Meen (this volume) discusses a related set of ‘complexity models’ yielding similar conclusions to those forwarded here. More modern sociological treatises closely related to collective socialisation also suggest thresholds, such as Wilson’s (1987) contention that as a critical mass of middle-class families leave the inner-city, low-income blacks left behind become isolated from the positive role models that the erstwhile dominant class offered. Economists also have developed several mathematical treatises involving collective socialisation effects in which thresholds often emerge as solutions to complex decision problems under certain assumptions (Akerlof 1980; Galster 1987; Brock & Durlauf, forthcoming). Seminal work in this vein has been produced by Schelling (1971; 1978), Schnare & MacRae (1975), Von Boventer (1978) and Taub et al. (1984). Variants are explored by Meen in Chapter 6. For deeper analyses, see Galster (1987), Grigsby et al. (1987), Rothenberg et al. (1991), Temkin & Rohe (1996) and Quercia & Galster (2000).

10 Access to Home Ownership in the United States: the Impact of Changing Perspectives on Constraints to Tenure Choice Joseph Gyourko

Introduction Access to home ownership is an important social and economic issue in many countries around the world. However, the factors determining that access vary across countries. This chapter reviews and discusses the marked change in emphasis in US research and policy circles on what constitutes the relevant budget constraint facing households making a tenure choice decision.1 Speciﬁcally, the American focus has shifted from permanent income to (non-human) current wealth as the binding constraint on access to home ownership. This change has had a number of important impacts that have inﬂuenced both the academic literature and US public policy surrounding home ownership. Given that Duncan Maclennan has played a prominent role in both the academic and policy worlds, this makes the topic particularly appropriate for inclusion in this volume. The ﬁrst impact of the change has been on research into tenure choice generally and has involved a recasting of the driving forces in theoretical models of tenure transition. Rather than permanent income and the relative user costs of ownership and renting driving the decision, a new characterisation has been adopted following the work of Jones (1995), in which current net worth and the asset price of the home drive the transition from renting to owning. This

Access to Home Ownership in the United States

173

new perspective is commented on more fully below. Empirical work by Jones (1989) and Linneman & Wachter (1989) conﬁrms that current non-human wealth is extremely important in explaining tenure differences across households and that it dominates permanent income as an explanatory factor. The second prominent impact of the new view of budget constraints in tenure choice has been an important change in how racial differentials in home ownership rates in America are viewed. Recent research using micro data ﬁnds that racial differentials in home ownership narrow considerably, or are eliminated altogether for the mean household, in speciﬁcations with good household wealth controls. This is particularly relevant in the US context, where large aggregate differences of 25 percentage points in home ownership propensities by race still exist. Building upon this work, Charles & Hurst (forthcoming) recently concluded that it is racial differences in mortgage application propensities that account for the bulk of the black/white difference in home ownership outcomes. Racial differences in income, in family structure, and in intergenerational transfers involving parental wealth help explain the mortgage applications gap. Hence, wealth – this time of parents – again plays a critical role. Naturally, this has important implications for policy makers interested in narrowing the aggregate racial gap in home ownership in the United States. The third major impact of the new perspective that current wealth is the primary binding constraint on the transition to home ownership has been on our understanding of the affordability of owner occupation. The dominance of wealth over income in empirical tenure choice studies strongly suggests that relaxing down-payment constraints, not policies that help households meet payment-to-income requirements (e.g., via lower interest rates), is what is needed to expand home ownership in America. Very low down-payment loans have proliferated in the United States recently, and the overall home ownership rate has increased. This is not to say that this is good policy, as the default experience for these loans is not yet known and has not been cycle tested. Nevertheless, the availability of low down-payment loans has been inﬂuenced by research that has helped change our perspective on what really constrains the transition from renting to home ownership. The remainder of this chapter is organised as follows. The next section introduces and discusses the new theoretical developments that are associated with the view that current wealth is the key constraint to tenure choice. It includes a brief discussion of the key initial empirical work testing the theory. This is followed by a section focused on the impacts of the new view on our understanding of racial differences in home ownership in the United States. Home ownership affordability issues are raised in the penultimate section. Finally, there is a brief summary and conclusion.

174

Housing Economics and Public Policy

Constraints to tenure choice – theory and empirical evidence The literature on tenure choice is a long and venerable one because of its importance in housing economics.2 The traditional model of tenure choice is one in which the rent/own decision is a function of the relative user costs of housing services obtained by owning versus renting and of permanent income, along with life cycle traits that reﬂect utility-based preferences for owner occupancy. Rosen (1979) provides an especially clear elucidation of this model in which households pick the tenure mode offering them the greatest utility. Utility from ownership is given by Vo = f(po, px, y)

(10.1)

where Vo is the indirect utility function conditional on owning, po is the user cost (or price per unit) of housing services associated with ownership, px is the price of all other goods (with x being the numeraire good), and y is permanent income. Analogously, Vr = f(pr, px, y)

(10.2)

where Vr is the indirect utility achieved conditional on renting and pr is the user cost (or price per unit) of housing services associated with renting. A household chooses to own if Vo>Vr. Empirical implementation of the model typically is done as part of a binary choice modelled in equation (10.3): V* = Vo – Vr = v(po/pr, y, z)

(10.3)

in which the price of housing services is put in terms of the numeraire good and z is included to capture household traits that might affect the utility associated with a particular tenure mode independent of the quantity of housing services actually consumed. Equation (10.3) then is expressed as a likelihood function, with a logit or probit model estimated. Underlying this structure is some uncertainty or imperfection that creates a preference for one tenure mode over another.3 The lumpiness of housing, incomplete rental markets, and non-neutral taxation have all been offered as candidates for why tenure preferences would arise.4 For our purposes, it is noteworthy that permanent income is the essential constraint on the tenure decision in this model. With housing being lumpy and its demand essentially determined from a multi-period optimisation of housing consumption, if

Access to Home Ownership in the United States

175

households are not liquidity constrained, then some threshold level of permanent income is required. Hence, in the absence of liquidity constrained households, higher permanent income is expected to be associated with a higher probability of owning. Jones (1995) provided a different perspective, arguing that tax non-neutralities in particular do not provide a solid foundation upon which to base a theory of tenure choice.5 Building on a series of primarily empirical studies indicating that many households were credit constrained,6 on data showing that a very large fraction of American and Canadian households past a certain age were owners and remained owners even in the face of substantial changes in incomes, interest rates, and the like, and on the theoretical implications of newly developed asset pricing models with illiquid durable goods,7 he suggested that there was a fundamental preference for ownership. In this view, stable households that are not especially mobile across housing markets are presumed to receive higher utility from ownership (holding constant the amount of housing services across tenure, of course). Who rents in this world? The answer is primarily younger, less permanently formed households, those households with little wealth available to make a down-payment, and those households with a relatively undiversiﬁed portfolio of highly leveraged, illiquid, risky assets. When do these households switch to owner occupancy? They change tenancy when the net change in their utility from switching is non-negative. In an inﬂuential paper on asset pricing by Grossman & Laroque (1990), the agent purchases a home only after a threshold amount of a riskless, liquid asset has been accumulated.8 Current liquid wealth, not the wealth associated with human capital embedded in permanent income, is the essential constraint to home ownership in this world. The difference in utility between owning and renting can be written as V* = v(w, ph, z)

(10.4)

where w is the net wealth of the household, ph is the stock price of housing (not the relative prices of a unit of the service ﬂows), and z captures demographics that reﬂect how permanently formed and/or mobile the household is. Relative user costs are omitted in the pure form of Jones’ (1995) model, as he argues that owned versus rental units are weak substitutes at best.9 This is a strong assumption that has not been fully accepted by some researchers in the ﬁeld. Indeed, much subsequent empirical work includes a control for the relative price of owning versus renting. This speciﬁcation issue has not affected the key ﬁnding that, in fact, wealth constraints dominate income constraints in terms of their practical impact.10

176

Housing Economics and Public Policy

While there are other interesting conceptual issues underlying the differences in these theoretical perspectives on tenure choice, we now turn to the recent empirical evidence, as it is that which motivates the changing views on race and home ownership and on housing affordability discussed in the introduction. While there were indications from earlier work that wealth constraints might be relevant to the tenure choice or tenure transition process,11 Jones (1989) and Linneman & Wachter (1989) provided direct evidence that down-payment requirements themselves were not only very important, but dominant factors empirically.12 Jones (1989) estimated the probability of owning as a function of household net worth (w), permanent income (yp, itself estimated), illiquid asset holdings (ia), and a set of demographic controls (z) as in equation (10.5) Pr(Own) = g(w, yp, ia, z)

(10.5)

Importantly, net worth (w) was speciﬁed as a series of dummy variables indicating by how much household net worth was less than or greater than house prices in the relevant market area. Illiquid asset holdings were included to test for the presence of a Grossman & Laroque (1990) and Plaut (1987) effect whereby households need to hedge housing price risk. The models were estimated on Canadian cross-sectional data, focusing on younger households.13 Linneman & Wachter (1989) pursued a similar approach using US data,14 but modelled their constraints based on the Federal National Mortgage Association (FNMA or Fannie Mae) and Federal Home Loan Mortgage Corporation (FHLMC or Freddie Mac) underwriting criteria for conventional loans. That is, they used the fact that these two agencies require mortgage insurance before they purchase any loan on a property with a loan-to-value (LTV) ratio of greater than 80% to guide their speciﬁcation of the borrower’s wealth constraint. This 20% equity down-payment requirement implies that the maximum home purchase price that satisﬁes the underwriting wealth criterion is ﬁve times household net worth. The authors then devised a series of variables to reﬂect the degree by which a household was wealth constrained. Similar measures were included to capture the extent to which the household was constrained by income-related underwriting requirements of Fannie Mae and Freddie Mac.15 Linneman & Wachter (1989) then estimated their models using the Federal Reserve Board’s 1977 Survey of Consumer Credit and its 1983 Survey of Consumer Finances. They restricted their sample to recent movers in both cross-sections. The results from both of these papers strongly indicated that wealth constraints dominated income constraints in terms of inﬂuencing tenure

Access to Home Ownership in the United States

177

status. Stated differently, the negative impact of being wealth constrained on the probability of ownership was far greater than that of being income constrained. A representative result from Linneman & Wachter (1989) is that the probability of ownership was about 32% lower among highly incomeconstrained households than among unconstrained households ceteris paribus, but this pales in signiﬁcance to the impact of being highly wealth constrained. For example, they report that highly wealth constrained households were 61% less likely to own than otherwise identical unconstrained households. This is an extremely large impact for a country in which just over 60% of all households own. Essentially, if an American household is wealth constrained as determined by standard underwriting criteria, the probability that it owns is quite low in absolute terms – even if it has all or many of the demographic traits such as being well-educated, over 35 years old, and a married head with minor children in the home that are typically associated with owner occupancy. In every sample examined in either study, the adverse impact of being highly wealth constrained exceeds that of being highly income constrained. Linneman & Wachter (1989) suggest that this is due to the fact that if a family cannot make the required down-payment on its desired home, its most viable alternative is to rent. However, this is not the case if the family is income constrained, but not wealth constrained. In this situation, the family can reduce its loan-to-value ratio below 80% in order to satisfy the payment-toincome criteria. This explanation that there is another alternative to renting when the household is income constrained, but not wealth constrained, is intuitively appealing. Other subsequent empirical work has conﬁrmed these initial results.16 This research has had signiﬁcant practical consequences; the increase in low down-payment mortgages, some with loan-to-value ratios in excess of 95%, in the mid- to late 1990s in the US is testimony to this.17 There is a broad (if informal) consensus among leading housing researchers and policy ofﬁcials that the substantial increase in home ownership rates in the 1990s from 60 to 68% at least partially reﬂects the easing of underlying wealth constraints.18 Thus, it is now clear to all that liqueﬁable wealth is the critical factor inﬂuencing access to home ownership – for Americans and Canadians, at least. And, it is equally obvious that this is due to equity-related needs associated with down-payment requirements. This immediately suggests that savings behaviour and savings ability, not merely current income and the level of interest rates, are important factors in determining who can own their residence. Moreover, the savings history of parents, not just younger households, could be important because of the extent of intergenerational

178

Housing Economics and Public Policy

transfers. As the next two sections highlight, these factors are crucial to our understanding of racial differences in ownership propensities in the United States and to our understanding of what really drives housing affordability.

Wealth constraints and racial differences in ownership in the United States Large aggregate racial differences in home ownership propensities have persisted over a long period in the United States. Decennial census data from 1970 showed that 61.2% of households with white heads owned versus 38.3% of households with black heads, so that the percentage of white households owning was about 1.6 times that of black households. By 1990, ownership rates had increased roughly proportionately for both races so that the percentage of white owners was still nearly 1.6 times that of black owners (69.4% versus 44.5%). These differences are of interest for at least three reasons: (a) the tax-favoured status of owner-occupied housing;19 (b) the important role that home equity plays in the creation and retention of household wealth;20 and (c) the social beneﬁts that many people attach to owning one’s home.21 The empirical work discussed in the previous section is relevant to this issue because it raises the possibility that controlling for wealth could substantially narrow the estimated racial difference in the probability of owning. Aggregate-level racial differences in net worth are known to be substantial. Using data from the 1994 wave of the Michigan Panel Study of Income Dynamics, Charles & Hurst (forthcoming) calculate that the mean wealth or net worth of black households is $43 365 versus $220 428 for white households. Skewness in the distribution of wealth certainly affects the means, but the wealth differential is large for the median black and white households, too. Speciﬁcally, the median black household’s net worth is $9435, while that for the median white household is $77 371.22 Thus, racial differences in wealth should be controlled for using micro data to see if this might help account for the large aggregate racial difference in the propensity to own. Results in Linneman & Wachter (1989) suggested that this might be very important. In fact, when evaluated at the mean values of the relevant parameters, Linneman & Wachter’s (1989) logistic function results indicate that there are no racial differences in ownership rates ceteris paribus. That is, once wealth is carefully controlled for, an otherwise typical household with a black head is no less likely to own than an otherwise observationally equivalent household with a white head.

Access to Home Ownership in the United States

179

Gyourko, Linneman & Wachter (1999) investigated more fully the possibility that the entire aggregate racial difference in ownership rates is due to racial differences in wealth that allow white people to meet down-payment requirements more easily. Using three cross-sections of the Federal Reserve Board wealth surveys from 1962, 1977, and 1983, they estimated a series of models that allowed them to measure racial differences in ownership probabilities for households away from the mean sample values. Wealth constraints in their paper were developed following the ‘regulatory rule’ perspective adopted in Linneman & Wachter (1989). Thus, a household is considered wealth constrained if its net worth is insufﬁcient to fund fully the down-payment implied by standard secondary market agency underwriting criteria for the purchase of loans. Given the 20% equity requirement imposed by Fannie Mae and Freddie Mac before mortgage insurance is needed, this means a household is categorised as constrained if the value of the home typically owned by other households with the same socio-economic trait set is more than ﬁve times net worth.23 In addition, this study included two other sets of variables to help control for the degree of wealth constraint (i.e., whether the household is barely constrained or hugely constrained in terms of its wealth shortfall, with unconstrained households having a zero shortfall) and for the degree of wealth ‘cushion’ (i.e., for unconstrained households, a measure of how much net worth exceeds the down-payment requirement of the household).24 Gyourko, Linneman & Wachter (1999) then estimated binomial logistic functions with all wealth constraint-related variables (and permanent income) interacted with a race dummy. Consistent with previous research, the results showed that wealth constraint status had a huge impact on the probability of owning. For an unconstrained household with a white head and other traits typical of an owner-occupier, the propensity to own was estimated to be 96% in the 1962 cross-section. Changing the wealth constraint status of that household is estimated to reduce the probability of ownership by 51 percentage points to 45%. For the analogous minority-headed household, the impact of being constrained is even larger at 80 percentage points (i.e., the probability of owning drops from 100% to 20%). However, the racial differences in implied ownership probabilities among unconstrained households are very small. The top panel of Table 10.1 reproduces results from Gyourko, Linneman & Wachter (1999) for the earliest and latest years they studied. Among unconstrained households, minorities are estimated to own at slightly higher propensities than their white counterparts,25 but the differences are quite small and the probabilities are quite high for all. This is the result implied in Linneman & Wachter (1989) when the regression ﬁndings are evaluated at sample means. Stated differently, there

180

Housing Economics and Public Policy

are no economically meaningful ceteris paribus differences in ownership rates among white and minority households who possess sufﬁcient wealth to meet down-payment and closing cost requirements associated with standard mortgage underwriting criteria. The bottom panel of Table 10.1 shows this is not the case for wealth-constrained households. Among this group, white households own at much higher rates than observationally equivalent minority households. The difference was 25 percentage points in the 1962 sample and 12 percentage points in the 1983 sample. These are not only large differences, but it is quite relevant to minority households as they are over-represented among constrained households. That is, these differentials apply to about one-third of the white households in the samples and to well over one-half the minority households. Thus, racial differences in wealth cannot account for all of the racial difference in ownership propensities. Moreover, the fact that many wealth-constrained households own, especially those where the head is white, suggests some type of systematic measurement error exists. Households categorised as constrained tend to have very low measured net worths – barely above zero. This indicates that households with low measured wealth have greater access to down-payment resources than is captured by household wealth in the Federal Reserve Board surveys. Other research suggests that intergenerational transfers from parents are a logical source of such unmeasured resources.26 Not only are such transfers Table 10.1 Predicted ownership probabilities and differential implied probabilities of ownership by race. 1962

Minority head White head Differential implied probability

Typical unconstrained head 1.00 0.96 0.04

Minority head White head Differential implied probability

Typical constrained head 0.20 0.45 –0.25

1983

0.97 0.91 0.06

0.17 0.29 –0.12

Source: Gyourko, Linneman & Wachter (1999), Table 6. See their paper for the details and deﬁnitions of the underlying reference households. Note: Gyourko, Linneman & Wachter (1999) used a white versus non-white characterisation of race. More detailed racial categories were experimented with (including a black versus non-black breakdown), but none of their essential results were changed.

Access to Home Ownership in the United States

181

large, there is no doubt that the parents of young minority households tend to have much lower net worths than the parents of young white households. While Gyourko, Linneman & Wachter (1999) speculate that racial differences in transfers from parents could account for the racial differences in ownership rates among constrained households, the fact is that their results cannot be thoroughly convincing as to the direction of causality because they arise from a cross-sectional analysis. The recent study by Charles & Hurst (forthcoming) addresses these issues using the University of Michigan’s Panel Survey of Income Dynamics to follow a sample of black and white renters over time. They analyse the factors causing the racial gap in housing transitions that occurs among their sample of renters. Importantly, they are able to investigate separately racial differences in the likelihood of applying for a mortgage and in the likelihood that a mortgage application is accepted or rejected. While there is a signiﬁcant racial difference in the probability of having a mortgage application rejected, the authors ﬁnd that this has very little explanatory power with respect to the much lower transition of black households to owner occupancy. Rather, the bulk of what they term the housing transitions gap is due to the fact that black people are far less likely to apply for mortgages in the ﬁrst place. Their investigation then shows that differences in income, family structure, and in parental transfers to help with down-payments are the primary reasons for the applications gap. Thus, parental wealth, not just younger household wealth, should be considered when modelling the budget constraint to the tenure choice decision. Just as the ﬁndings in Jones (1989, 1995) and Linneman & Wachter (1989) raised an important policy question regarding what was necessary to expand home ownership rates, those in Charles & Hurst (forthcoming) raise the same issue with respect to black households. There is a huge literature on discrimination in the mortgage market,27 and while Charles & Hurst (forthcoming) do ﬁnd evidence that lenders do not treat black and white households equally, it is wealth – either their own or of their parents – that appears to be so constraining for minorities. The evidence is more convincing now because it has shown up in panel data that can be exploited to identify more carefully the direction of causality. Thus, ﬁnding ways of relaxing those constraints should be a priority for those who wish to address the aggregate racial difference in ownership probabilities. While discrimination in the lending market certainly should be stamped out wherever and whenever possible, policy makers and households should understand that that alone will do little to narrow the racial gap in ownership.

182

Housing Economics and Public Policy

Implications for the affordability debate In the 1980s, the affordability of single-family owner-occupied housing joined traditional issues such as housing quality and racial discrimination as a focus of housing policy debates in the United States. While an ageing population led many to predict rising home ownership rates during the decade, the aggregate ownership rate actually declined by 1 percentage point, marking a reversal of a trend dating back to the end of the Second World War.28 While concern over the affordability of owner occupation for lower- and middle-class households led to a national housing bill entitled the National Affordable Housing Act of 1990, the fact was (and still is) that what is meant by housing affordability is not precisely deﬁned. However, as Linneman & Megbolugbe (1992) have noted, how affordability is deﬁned can have important policy consequences. Prior to the research ﬁndings on wealth constraining the transition to owner occupancy, interest rate policy was seen as the key to the affordability issue. This is implicit in the most widely known affordability index in the United States, which is published by the National Association of Realtors (NAR). This index is constructed such that an index value of 100 implies that the median income family qualiﬁes for the median value home. Because interest rates are much more variable than income, changes in the NAR Housing Affordability Index over time primarily reﬂect changes in interest rates. This is evident in Fig. 10.1, reproduced from Gyourko & Tracy (1999). Whenever long-term mortgage rates are low, the NAR series signals excellent affordability conditions (and vice versa). Given the continued low long-term inter1.10

18 30-Year Fixed Rate Mortgage Rate [ right axis => ]

1.00

16 Composite NAR Scale [<= left axis]

0.90

14

0.80

12

0.70

10

0.60

8

0.50

6

0.40 1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

4 1998

Year

Fig. 10.1 National Association of Realtors (NAR) housing affordability indices. Source: Gyourko & Tracy (1999), Chart 1.

Access to Home Ownership in the United States

183

est rates in America, this series suggests that housing is now more affordable than at any time since 1974. However, various economic changes including a substantial increase in the dispersion of income-by-skill group cast doubt as to whether it is sensible to focus solely on the affordability of the median-priced home. While lower interest rates certainly do reduce the income necessary to purchase a home, they do not directly reduce the down-payment-related requirement. Thus, consistent with the implications of the research cited above, the levels and growth rates of savings and incomes (in addition to house prices and interest rates) are key components of housing affordability. Two articles by Gyourko & Linneman (1993) and Gyourko & Tracy (1999) try to use the insights of the new view of wealth constraints as the key to tenure choice to move beyond the NAR Housing Affordability Index. Using decennial census and American Housing Survey (AHS) data, Gyourko & Linneman (1993) tried to answer a simple question: is a home of a given quality from 10–15 years ago more affordable or less affordable today to a household similarly situated to the one that occupied the home then? To perform their analysis, they constructed unadjusted and quality-adjusted price indices for ﬁve types of homes. The latter series were developed using standard hedonic techniques described in their paper. A ﬁrst point to note from that research is that real house price appreciation varies signiﬁcantly over time and across the house price distribution. That is, cheap and expensive homes have fared far differently over time in terms of their price growth. This is illustrated in Table 10.2. These data are for raw, unadjusted prices that do not hold constant the quality of the homes in the Table 10.2 Real house price appreciation – US national data aggregate and average per annum (in parentheses) ﬁgures. Time period Percentile of the house price distribution 1960–74

1974–81

1981–89

10th

54.1 (6.4) 39.6 (4.9) 17.3 (2.3) 19.2 (2.5) 41.2 (5.1)

–38.1 (–5.8) –28.7 (–4.1) –16.8 (–2.3) –1.5 (–0.0) –2.0 (–0.0)

25th 50th 75th 90th

35.1 (2.2) 48.8 (2.9) 44.5 (2.7) 46.5 (2.8) 49.2 (2.9)

Source: Gyourko & Linneman (1993), Table 4. Calculations made from decennial census and American Housing Survey data.

184

Housing Economics and Public Policy

10th, 25th, 50th, 75th, and 90th percentiles of the house price distribution. It is clear that the 1980s witnessed a dramatic reversal in the earlier pattern of consistently positive real house price appreciation in the United States, with the percentage declines being especially large among lower valued homes. In fact, the drops in real value for the cheaper homes at or below the 50th percentile were so large that by 1989 their real prices had reverted to 1974 levels. The constant quality price series estimated by Gyourko & Linneman (1993) then tell a different story.29 Table 10.3 reproduces results that compare constant quality and unadjusted price appreciation for the different types of homes. Note that constant quality real price growth is signiﬁcantly greater than the appreciation in actual (or unadjusted) prices for lower quality bundles.30 Not only does this indicate that the raw price series may seriously misrepresent the affordability of a speciﬁc low- or moderate-quality unit over time, it also suggests that there may have been a serious erosion of housing quality among lower priced homes. Whether this was due to changes in demand fundamentals or to the inability of many lower income households to maintain their homes adequately after expending most of their wealth to purchase them is unknown. At a minimum, it raises the spectre of savings and wealth once again, importantly inﬂuencing the ability to own and to protect the equity in one’s home. Table 10.3 also shows that the pattern of actual versus constant quality prices is very different for more expensive and, presumably, higher quality homes. At the upper end of the price distribution, appreciation in the unadjusted series exceeds that for the constant quality series, suggesting that quality improvements can help account for the huge price growth. Consistent with much research in labour economics, Gyourko & Linneman (1993) also showed that real wages of the least-skilled workers (deﬁned as those without a high school degree) had actually declined between 1960–1989. Combined, the constant quality price data and the earnings data suggested that households headed by low-skilled workers could afford to own singlefamily housing at the end of the 1980s only by having two earners or by reducing the quality of the home consumed (or both). Since there are limits to how far down the quality spectrum one can shift from the 10th or 25th percentile home, it seems likely that this is where the affordability problem has become most serious. These households simply cannot afford to save for a down-payment, and the data are at least suggestive that they may be allowing their housing capital to depreciate if they are able to become owners. Gyourko & Linneman (1993) concluded that there was not a similar change in affordability conditions for the typical occupant of a home at or above the median quality. Wages and salaries for these earners were stagnant or grow-

Table 10.3 Constant-quality and unadjusted price appreciation: homes and trait bundles from various percentiles of the house price distribution. 1974 trait bundles 1974–1989 (%)

1974–peak (%)

Peak–1989 (%)

28.1 1.7

54.6 6.4

–17.2 –2.3

13.1 0.8

20.2 4.7

13.7 0.9

1974–1989 (%)

1974–peak (%)

Peak–1989 (%)

–4.7 –0.3

54.1 6.4

–38.1 –5.8

–5.9 –0.6

–0.5 –0.0

39.6 4.9

–28.7 –4.1

22.8 5.3

–7.5 –0.7

–2.5 –0.2

22.6 4.2

–20.4 –2.3

4.4 0.3

29.2 6.6

–19.2 –1.9

17.4 1.1

48.3 10.4

–20.1 –2.0

4.4 0.3

28.7 6.5

–18.9 –1.9

38.3 2.2

41.7 7.2

–2.4 –0.2

Source: Gyourko & Linneman (1993), Table 6. Calculations made by the authors based on 1974–1989 American Housing Surveys. Note: Calculations also made using trait bundles from 1989. The implications for quality changes are unaffected. See Gyourko & Linneman (1993) for the details.

Access to Home Ownership in the United States

10th percentile Total appreciation Avg. annual appreciation 25th percentile Total appreciation Avg. annual appreciation 50th percentile Total appreciation Avg. annual appreciation 75th percentile Total appreciation Avg. annual appreciation 90th percentile Total appreciation Avg. annual appreciation

Unadjusted prices

185

186

Housing Economics and Public Policy

ing, and the real prices of high quality, constant-quality housing bundles barely rose. Hence, the moderate or high quality home from the mid-1970s still is affordable to these households, even if the higher quality stock that has been built since then is not. Gyourko & Tracy (1999) further examined the affordability issue by updating the analysis with data from the 1990s and by employing a quantile regression approach in addition to the mean regression approach standard in the literature. The quantile regression approach uses methodology similar to the mean regression approach, but relaxes the restriction that only average trait prices are used to construct the constant-quality price indices. That is, each individual price index (e.g. the one for the 25th percentile home) is constructed using its own trait prices. More speciﬁcally, the trait prices for the 25th percentile are selected so that 75% of actual home prices are higher than what one would predict based on the house traits and on the 25th percentile trait prices. In addition, 25% of actual house prices are lower than what one would predict based on the house traits and on the 25th percentile trait prices. Thus, if a trait such as bathrooms tends to contribute relatively more value to high-quality homes than to low-quality homes, then this will show up as differences between the quantile-speciﬁc price for bathrooms at the upper and lower ends of the house quality distribution.31 Figures 10.2, 10.3, and 10.4 reproduce Gyourko & Tracy’s (1999) ﬁndings for the raw price series, for the constant quality series using the mean regression approach, and for the constant quality series using the quantile regression 2.00

90th Percentile

1.80

1.60 75th Percentile 1.40

1.20

50th Percentile

1.00

0.80

10th Percentile 25th Percentile

0.60 1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

Fig. 10.2 Real house price distribution – unadjusted series. Source: Gyourko & Tracy (1999), Chart 5; underlying data are from the AHS national series.

Access to Home Ownership in the United States

187

1.40

10th Percentile

1.30

1.20

75th Percentile

1.10

50th Percentile

1.00

90th Percentile

25th Percentile

0.90

0.80 1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

Fig. 10.3 Constant quality house price indices mean regression method. Source: Gyourko & Tracy (1999), Chart 6; series estimated by authors using the AHS national series.

1.80 1.70 90th Percentile 1.60 1.50 1.40

75th Percentile

1.30

10th Percentile

1.20 1.10 1.00 50th Percentile 0.90 0.80 1974

25th Percentile

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

Fig. 10.4 Constant quality price indices quantile regression method. Source: Gyourko & Tracy (1999), Chart 7; series estimated by authors using the AHS national series.

approach. While the latter two approaches share many common features, there are some important differences in results in the upper and lower tails of the price distribution. In the upper tail, the quantile approach suggests there was more rapid real price growth between 1974 and 1997. While the average hedonic measure indicates that the 90th percentile constant quality price was only 1% higher in 1997 than in 1974, the quantile hedonic measure suggests the increase was 31%. Figure 10.2 indicates that the unadjusted series

188

Housing Economics and Public Policy

for the 90th percentile home was 35% higher in 1997 than in 1974. The implied increase in quality among higher-end homes looks to be considerably smaller when estimated using the quantile regression approach. At the bottom of the price distribution, the average hedonic price index (Fig. 10.3) suggests that a constant quality house at the 10th percentile was 33% more expensive in 1997 than in 1974. The quantile speciﬁc index plots a similar, but less stark picture, as the analogous appreciation measure was 20%. The fact that the 10th percentile unadjusted series is well below both constant quality series suggests that average quality has worsened at the bottom of the house price distribution, with the magnitude varying by estimation strategy.32 In sum this research, and the work discussed above that conﬁrms the importance of wealth constraints, should lead us away from reliance on affordability measures that focus on the median household or which rely exclusively on income or interest rates in their composition. While the NAR’s affordability index suggests that home ownership opportunities have improved steadily throughout much of the past two decades, other evidence suggests scepticism is in order in this regard, especially as one moves down from the median home or the median earner. While the quantile regression results of Gyourko & Tracy (1999) indicate the constant quality price of the 10th percentile home did not rise as much as implied by the mean regression approach used in Gyourko & Linneman (1993), both papers conclude real constant quality prices of that lower quality home have risen. And the data from the labour market continues to show a widening dispersion in wages, with low skilled workers performing the worst in absolute and relative terms. It is difﬁcult to imagine that wealth constraints will not continue to be binding for such households. If we care about such households owning, then it must be recognised that series driven by interest rates will not give an accurate picture of the affordability conditions they face.

Summary and conclusion Much has been learned over the last decade about the constraints involved in the transition from renting to owning. New models and abundant empirical work show that wealth constraints are the critical hurdle facing most households. This new insight from the academic literature has important practical implications for those who believe that expanding home ownership will bring valuable social and economic beneﬁts to new owners. With respect to basic affordability, it is clear that low interest rates alone cannot make ownership affordable, particularly to low wealth households.

Access to Home Ownership in the United States

189

It is not at all clear that very low down-payment loans of the type that has proliferated in the United States in the latter half of the 1990s will turn out to be a good idea – ﬁnancially or socially. Work in a British context (Maclennan et al. 1997) strongly suggests that there are costs in a downturn associated with extending low down-payment loans to households in periods of strong economic growth. While the American and British markets are different enough to rule out simple extrapolations, such work is cautionary. However, it is clear that something like this is needed to surmount the most basic of affordability constraints. In addition, recent research suggests that racial differences in wealth play a large role in accounting for the wide aggregate racial differences in home ownership that exist in the United States. In this case, parental wealth is an important factor, not just that of the younger household wishing to become an owner. At a minimum, policy makers in America should be aware that an end to alleged discrimination in the mortgage market will not substantially narrow the racial gap in home ownership, in the absence of a policy that deals with the very uneven racial distribution of liquid wealth. Finally, it is clear that future research should not focus so squarely on the median household. Increasing wage dispersion in the labour market, among other factors, has acted to make affordability conditions vary across the income distribution. Different estimation techniques should be used to provide better insight into just how affordability varies across different types of households.

Acknowledgements I particularly appreciate the comments of Kenneth Gibb, Peter Linneman, and Tony O’Sullivan on previous drafts. Naturally, they are not responsible for any remaining errors.

Notes (1)

(2)

While accessibility to home ownership is an important international issue, this chapter focuses on the US situation and literature. Covering multiple countries is not feasible without a material increase in paper length because key country-speciﬁc factors such as mortgage market regulations and racial effects differ so much across nations. The literature is voluminous and an exhaustive bibliography will not be attempted here. An excellent understanding of the key issues can be obtained from the following selected articles: Artle & Varaiya (1978), Weiss (1978), Rosen

190

(3)

(4)

(5)

(6) (7) (8) (9)

(10)

(11)

(12)

(13) (14)

(15)

Housing Economics and Public Policy (1979), Hendershott (1980), Brueckner (1986), Poterba (1991), and Henderson & Ioannides (1983). See Arnott’s (1987) review for the details. With perfect knowledge and complete markets, neutral taxation, divisible housing and other assets, and no transactions costs, households would be indifferent between owning and renting. Because housing is heavily subsidised under the US tax code, tax nonneutralities are particularly well studied in the American literature. Rosen (1979) is the canonical example of this strand of research. The interested reader should see Follain & Ling (1988) and the citations therein for a review. Jones (1995) noted that the empirical literature is ambiguous at best regarding the hypothesis that the probability of ownership is decreasing with relative user costs (po/pr). Linneman (1985) also provides strong evidence that the tax sorting implications of the tax-based models are not borne out in reality. In particular, see Jones (1989), Linneman & Wachter (1989), Haurin (1991), and Duca & Rosenthal (1994). See Grossman & Laroque (1990). Households need to hedge housing price risk in these models. See Plaut (1987) for an early version of this insight applied to the tenure transition decision. This reﬂects his underlying assumption that there exists a fundamental preference for ownership. In addition, it must be the case that one cannot expect a higher return to owner-occupied versus rental housing (regardless of one’s tax bracket). See Jones (1995) for the details. In Jones (1995), the implied wealth threshold in equation (10.4) is endogenously determined. Other researchers have proposed models with exogenous constraints, typically in the form of down-payment requirements. See Artle & Varaiya (1978), Brueckner (1986), and Engelhardt (1994b). For example, Dynarski & Sheffrin’s (1985) ﬁnding that transitory income affects the probability of ownership hints that this complements net worth and that there is more than permanent income to the constraint story. Similarly, Henderson & Ioannides’s (1987) conclusion that the steepness of one’s permanent income path affects the probability of owning also suggests that permanent income alone is not all that is relevant. Credit should also be given to Bossons (1978), who Jones (1989, 1995) drew on for inspiration, and who provided the very ﬁrst empirical evidence of which I am aware that household wealth strongly inﬂuenced the choice of tenure mode. Speciﬁcally, the micro data were from Statistics Canada’s 1977 and 1984 Surveys of Consumer Finances (SCF). Although they included a user cost measure and therefore did not estimate the pure tenure transition model advocated by Jones in his 1989 or 1995 papers, as I have already indicated, this is not signiﬁcant for our current purposes. Fannie Mae and Freddie Mac underwriting requirements stated that annual mortgage payments needed to be no more than 28% of the borrower’s annual family income. See Linneman & Wachter (1989) for the details.

Access to Home Ownership in the United States (16) (17)

(18)

(19) (20)

(21)

(22) (23)

(24)

(25)

(26)

191

For example, see Duca & Rosenthal (1994) and Gyourko, Linneman & Wachter (1999). Low down-payment loans ease the equity or wealth constraint while increasing the severity of the income constraint, as special mortgage insurance or higher interest rates (or both) typically apply on these loans. Hence, the income burden is more severe. That said, this is not all due to the spread of low down-payment mortgages. Rising real wages and increased wealth creation during the 1990s also made it easier for households to save for the 20% down-payment on standard mortgages that do not require mortgage insurance or bear a special risk premium. Poterba (1991) estimates that tax beneﬁts lower the user cost of owning by about 15% for the highest income (and highest tax bracket) households. Recent research by Tracy et al. (1999) shows even with the great boom in equities in the US in the 1990s, home equity still constitutes virtually all of household wealth for the vast majority of households, including the median household. See Green & White (1997) and DiPasquale & Glaeser (1999). They ﬁnd that owning one’s home is associated with better social outcomes for children and enhanced political and community involvement. That said, Oswald (1997b, 1999) believes there is a downside to this social capital, primarily in terms of higher unemployment. See Charles & Hurst (forthcoming) for the details behind their calculations. We abstract here from a host of empirical issues such as closing costs and the estimation of the ‘desired’ home. See Gyourko, Linneman & Wachter (1999) for the details. Consistent with the results in Jones (1989) and Linneman & Wachter (1989), this study found non-linear effects associated with the extent to which a household was constrained. Basically, once the household is quite constrained, being constrained by another dollar has little further impact on tenure outcomes. The results for the wealth cushion variables were consistent with Grossman & Laroque (1990) effects, as households with net worths just barely above their implied down-payment requirements had lower probabilities of owning compared to households with more substantial wealth cushions. Thus, the data are consistent with the implication that households prefer not to be cornered in illiquid and risky housing. The slightly higher estimated ownership propensities among unconstrained minorities probably reﬂect sample selection bias. Particularly in 1962, but in later years too, widespread discrimination in labour markets prevented many minorities from amassing enough wealth to be unconstrained. Minority households able to surmount that discrimination probably possess unobserved traits making them disproportionately likely to own. Research on intergenerational transfers is part of a broader literature on savings behaviour. See Kotlikoff (1988) for a review. In the housing context, see Engelhardt (1994a, 1994b, 1995) and Engelhardt & Mayer (1994, 1995) for analysis of how savings behaviour interacts with parental gifts and downpayment constraints.

192 (27)

(28)

(29)

(30)

(31) (32)

Housing Economics and Public Policy Examples of recent work include that by Munnell et al. (1996) on mortgage acceptance and rejection and by Berkovec et al. (1994) on default. Quigley (1996) provides an excellent review of the literature accompanying the debate on whether or not there is discrimination. Lastly, Kain & Quigley (1972a) is the pioneering work on racial differences in ownership and location outcomes. In the US, affordability concerns tend to be focused on home ownership. In many other countries, the focus is more on rental affordability (e.g., see Maclennan & Williams 1990a,b, and Maclennan, Gibb & More 1990). In America, rental affordability is often viewed as part of a broader set of welfarerelated issues, consideration of which would take us too far aﬁeld from the issues of primary concern in this chapter. Ten quality bundles were priced via hedonic regressions. Five trait bundles from 1974 were priced going forward, and ﬁve 1989 trait bundles were priced going backward in time. Traits controlled for were the number of bathrooms, the number of bedrooms, the number of other rooms, whether the unit was detached, whether there was a garage, whether there was a cellar, the type of heating system, whether there was central air conditioning, neighbourhood quality, overall house quality, whether the home was in the central city of its metropolitan area, and house age entered as a quadratic. The speciﬁc bundles themselves were determined by the traits typical of homes in the 10th, 25th, 50th, 75th, and 90th percentiles of the single-family house price distribution in 1974 and 1989, respectively. See Gyourko & Linneman (1993, Appendix Tables 1 and 2) for the details. The 1974 and 1989 bundles associated with the lowest quality homes appreciated by 28.1% and 33.2%, respectively, over the 1974–1989 period, while the unadjusted price for the home in the 10th percentile of the price distribution actually declined by 4.7% (in real terms). The speciﬁcation itself was identical to that in Gyourko & Linneman (1993). See Note 29 for the details. All that said, one still cannot simply conclude that quality changes must underpin any differences between unadjusted and constant quality price growth rates. The average hedonic method may miss demand-induced price changes, but the quantile hedonic method may price up quality changes in addition to demand-induced price effects. See Gyourko & Tracy (1999) for the details. The average and quantile hedonic methods may provide a way to bound the true unobserved constant quality price index.

11 Planning Regulation and Housing Supply in a Market System Glen Bramley Introduction A basic contention of this chapter is that housing supply matters. Housing supply structures and mechanisms have a number of important economic effects on, among other things, macroeconomic cycles, wealth accumulation, overall welfare and its distribution, economic competitiveness, regional labour market imbalance, urban vitality and environmental performance. A few British-based examples serve to underline this general contention.

•

New housing supply is relatively inelastic in Britain, and this exacerbates price volatility across the economic cycle and gross price disparities between regions. These price effects have major implications for the distribution of wealth and probably serve to further destabilise the macroeconomy (Bramley et al. 1995; Meen 1996, 1999b).

•

Instability of the housing market, resulting partly from inelastic supply, probably exacerbates the technical conservatism of residential construction and its mediocre quality record (e.g. because of reliance on subcontracting, undersupply of training) (Ball 1996a).

•

Competitive behaviour in the face of regulation may lead to deadweight costs of developer activities designed to gain planning permission (rentseeking), and there are dangers of impropriety/corruption in the system; the way in which planning and development control are applied may serve to increase delay and uncertainty (Evans 1991; Cheshire & Leven 1982; School of Planning & Housing 2001).

•

Local political control over planning may create systematic biases in land supply for different activities, so affecting overall competitive performance (Lambert & Bramley 2002).

194

Housing Economics and Public Policy

•

Mechanisms for infrastructure provision and ﬁnancing, in an era of public expenditure constraint when most development is privately promoted, may be incoherent and ineffective, giving rise to delays and inequities (Healey et al. 1993; Monk & Whitehead 2000; Department of Transport, Local Government and the Regions 2001).

•

Planning may be used to support urban renaissance and vitality, but it may also frustrate it (Department of the Environment 1999b; LlewellynDavies 1994).

However, it seems that the supply side of the housing market receives limited attention compared with that lavished on the demand side. How can we explain this imbalance? Is it really the case that there are no interesting questions about the supply of housing? Or is it rather that our standard economic toolkit does not help very much with addressing such questions? At a theoretical level, for example, the assumptions of competitive behaviour may assume away the problems. In practice, to address the major issues may involve engaging with messy institutional or political processes, requiring a retreat from the preferred deductive modelling paradigm. Even if we can describe some of these processes, we may not be able to measure them. In simple terms, housing supply may be decomposed into three main dimensions:

•

construction of new dwellings;

•

management of existing housing, including issues of ownership, access, pricing and ﬁnancing; and

•

maintenance of housing, including repair and up- (or down-) grading of the stock.

This chapter addresses the ﬁrst of these dimensions, presenting some empirical work undertaken using English data. The main focus of attention is upon land use regulation and its impact on housing supply and thence on the housing market. However, understanding this impact entails assumptions about both the structure of the industry and the structure of regulation, including its linkage with infrastructure and ﬁscal issues. The next section brieﬂy reviews past economic research on new housing supply. This draws out the particular signiﬁcance of land supply and the key role of land use planning in inﬂuencing this process. The third section discusses planning land use regulation in general terms, describing the op-

Planning Regulation and Housing Supply in a Market System

195

eration of the British system and drawing out some contrasts with systems in other countries. This section also reviews previous efforts at measuring the economic impact of planning. The following section sets the scene for subsequent empirical analysis, describing the inter-urban panel dataset used and commenting on certain aspects of the study, particularly its geographical framework and its theoretical underpinnings. The ﬁfth section presents the empirical modelling results. The elasticities derived are compared with previous studies, and in the concluding section brief comments are made on the signiﬁcance of these ﬁndings for policy.

Researching housing supply Econometric models of housing have hitherto followed two traditional approaches: (a) macro and regional time series; (b) hedonic house price models and related urban models. The former have become increasingly sophisticated in their treatment of dynamics and trends over time, including the application of cointegration techniques (Meen 1996, 1998, 1999b; Guissani & Hadjimatheou 1991; Muellbauer & Murphy 1997; Munro & Tu 1997). While the major emphasis has been on cycles in demand and house prices, there have been quite sophisticated models of new building supply and/or investment in this tradition (Tsoukis & Westaway 1991; Ball 1996b). However, up until now this type of research has lacked any data to get a handle on land supply and the inﬂuence of planning policy and practice. The latter type of research, the dominant tradition in US urban economics, typically involves cross-sectional analysis at either individual property level or at small area level. On the whole, most of the thrust of this work has concerned questions about demand, including measuring the demand for different attributes within the housing bundle, and measuring the demand/ willingness to pay for environmental beneﬁts (e.g. Michaels & Smith 1990). Cheshire & Sheppard (1989, 1997) explore this approach to measuring some of the beneﬁts of planning regulation, for example, urban open space protection. A good part of this kind of work has been concerned with the emergence or persistence of ‘sub-markets’ (Schnare & Struyk 1976; Goodman 1981; Adair et al. 1996; Maclennan & Tu 1996), and part of the reason for the existence of these may be supply constraints (as recognised by Maclennan 1977b, 1982; Ball & Kirwan 1977). Another tradition within urban economics has been the analysis of housing supply in terms of the intensity of land use (following Muth 1969; see reviews in Bartlett 1991; Bramley et al. 1995, Chapters 2 & 8). However, there seems to have been relatively little intraurban and cross-sectional analysis of new build supply itself.

196

Housing Economics and Public Policy

There is also a signiﬁcant qualitative research tradition in this ﬁeld, for example involving key actor interviews and case studies (often with a longitudinal element). Such work may be informed by standard microeconomic theory, by emerging perspectives including public choice and institutional economics, or by perspectives from other disciplines. Examples of work within this tradition include Monk & Whitehead (2000), Rydin (1985, 1986), Short et al. (1986) and Barlow (1993). Other work gives insight into the motivation and behaviour of landowners (Adams & May 1992; Adams et al. 1999), as well as the planners and housebuilders who tend to be the primary focus. Organisations with a policy interest in the ﬁeld support work that provides a combination of monitoring data, local case studies and critical commentary (Council for the Protection of Rural England 2001; Bate 1999; Industrial Systems Research 1999; Joseph Rowntree Foundation 1994). One key challenge is the development of systematic conceptualisation and measurement of planning/regulatory restriction (Bramley 1998; Malpezzi 1996; Needham & Lie 1994). There is clearly potential for more sophisticated theoretical treatment of local planning regulation, possibly involving the techniques of game theory. Some US ‘local public choice’ literature treats planning (‘zoning’) as one of the key strategies available to local jurisdictions engaged in ‘ﬁscal competition’ for businesses and (some kinds of) residents (Tiebout 1956; Hamilton 1975; Fischel 1990). One of the paradoxes of researching housing supply is that it cannot actually be separated from demand. Land is not a factor like other factors, and with its spatial ﬁxity and restricted supply, land cost reﬂects demand-based value; in other words, land cost is endogenous (Anas & Arnott 1993a; Bartlett 1991). Furthermore, with signiﬁcant lags in the production process (Maclennan 1982), the derived demand for housing land or new construction is subject to expectations and uncertainty about a housing market that is itself highly unstable (Bramley 1989; Bramley et al. 1995).

Planning and regulation of land use The differing nature of planning/regulatory regimes As stated, the main purpose of this chapter is to look at practical attempts to model empirically the impact of land use regulation, or ‘planning’ as it is generally known in Britain, on new housing supply and the housing market. As a preliminary to that it is perhaps useful to clarify what these terms mean, and to draw attention to important differences in the nature of land use planning and regulation between Britain and other English-speaking countries.

Planning Regulation and Housing Supply in a Market System

197

Land use regulation is a legal or administrative system that prescribes what type of use is permitted on any given parcel of land. ‘Use’ encompasses both the type of activity accommodated (e.g. housing, industry, retailing) and, to some degree, the type, scale and form of buildings that may accommodate that use. Such regulation comes into play most often when landowners seek to change the use of a piece of land; a change of use constitutes ‘development’. There are perhaps two main reasons why this form of regulation comes to be associated with ‘planning’. One reason is that a common and natural way to express current or permitted patterns of land use is in the form of a map (i.e. literally, a plan, in the sense of blueprint). A second reason is that decisions about future land uses/developments to be permitted are often made in the context of a forward looking, comprehensive policy procedure and document, also generally known as a plan (but this time emphasising the time dimension and the notion of intentionality). It should not be forgotten that planning as just deﬁned is not exclusively a public function. However, in general, advanced societies have found it necessary or desirable to interpose a public legal and administrative basis for land use regulation and planning. In the USA and many other countries this primarily takes the form of ‘zoning’. Under this approach the territory is divided up into zones within each of which certain types of land use are permitted subject to certain parameters, for example minimum lot sizes and building lines for housing. The essence of this approach is that the landowner has the right to develop within these parameters (and the corresponding right to expect that neighbouring plots will not be developed in a ‘nonconforming’ fashion). (Cullingworth 1997a provides a transatlantic perspective on the system as it has actually evolved). This approach contrasts strongly with the British system, under which since 1947 land use rights have been ‘nationalised’ and there is no automatic presumption in favour of development. Any signiﬁcant development requires permission from the local planning authority, which has considerable discretion in decision-making (Grant 1992). One of the interesting developments of the 1990s has been an attempt to make the British system more ‘plan-led’, so that there is a general presumption that there should be operative, sitespeciﬁc local plans in place, and that a site ‘allocated’ for a particular use has a presumption in favour of development for that use. This can be seen as a move towards more of a zoning approach, although commentators are quick to qualify any such conclusion (Tewdyr-Jones 1996). Formally the British approach entails a hierarchical system, ranging from national policy guidance, regional planning guidance (in England), structure plans (for county or sub-regional areas), and local plans. The apparently com-

198

Housing Economics and Public Policy

prehensive character of this policy framework and forward planning system distinguishes it from that found in many other countries. In principle, local development control decisions should be determined on the basis of a rational comprehensive system that works out the local spatial implications of general policies in a way that is locally sensitive and democratically accountable. The system as it operates in practice falls somewhat short of this ideal. Shortcomings include the tendency for plans to be slow to be completed or updated, so that the currently available plans may not reﬂect current information or policies (School of Planning & Housing 2001), and hence there remains considerable scope for discretion. Most of the delays that characterise development control decisions, particularly on larger developments, result from the negotiations that such discretion allows. These negotiations are often concerned with infrastructure provision, which increasingly developers are expected to pay for. Currently the British planning system is under substantial scrutiny and review, reﬂecting a range of concerns about speed of decision-making, responsiveness to economic needs, and whether it performs its longer term spatial directing role effectively. There are contradictions between some of the desiderata which different commentators and lobbies put forward for planning – more certainty versus more ﬂexibility, more economic responsiveness versus more environmental sustainability, more local democracy and participation versus more rationality and efﬁciency. Thus it is difﬁcult to predict how the system will evolve in practice. However, the hitherto very different systems of Britain and North America may be converging to some extent. This is particularly apparent now that ‘growth controls’ and ‘smart urban growth’ have become important features of the American scene.

The economic impact of planning and regulation There is broad agreement among economists about the nature and direction of the effects of planning and land use regulation on the housing market, based on economic theory. Planning/regulation that signiﬁcantly restricts the supply of land for new housing development is likely to reduce supply elasticities, increase land and house prices, and increase housing densities. However, there are theoretical grounds for differences of view about (a) the properties of a notional ‘no planning’ regime and (b) the welfare interpretation of these impacts. On the ﬁrst point, for example, there are grounds for doubting whether supply elasticities would be inﬁnitely high in the absence of planning, owing to issues of accessibility, infrastructure availability and uncertainty (see Bramley 1989; Evans 1983; Wiltshaw 1985; Neutze 1987). On the second point, planning/regulation may create positive environmen-

Planning Regulation and Housing Supply in a Market System

199

tal amenity effects that account for at least some of the apparently higher prices associated with the regime (Cheshire & Sheppard 1989, 1997; Monk et al. 1991). The more important arena for debate is around the empirical magnitude of the various effects postulated. This shifts the emphasis to issues of empirical measurement and modelling, including both understanding and measuring planning regulations themselves, modelling supply responses, and modelling housing market adjustment to these effects over time and space. This applied empirical focus is wholly in the spirit of Maclennan’s contributions to housing economics and has informed the present author’s contributions, including the empirical work reported later in this chapter. Lack of space prevents a full review of existing empirical literature, but it is useful to highlight the main points emerging so far, drawing on a number of useful reviews including Fischel (1990), Monk et al. (1991), Podogzinski & Sass (1991) and School of Planning & Housing (2001). US studies have mainly employed hedonic house price modelling to examine the effects of zoning and the recent proliferation of growth controls. There is a surprising lack of unanimity across different studies, with some appearing to show that land use controls are not an effective constraint, whilst others argue that they are and that they do push up housing and developed land prices. The impact depends in part on the degree of competition between jurisdictions and the extent to which controls are generalised. Fischel (1990) argues that for political/legal reasons growth controls tend to go ‘too far’ and create net welfare costs for society in terms of excessive sprawl (leapfrogging beyond areas of control) and increased commuting. Cheshire & Sheppard (1989, 1997) is the most directly comparable British work, that applies a sophisticated intra-urban model to measure the impact of planning controls and planned amenity provision in a limited number of cities. Their general conclusion is that British planning control, by containing urban extension, tends to increase house prices moderately but has its main (adverse) welfare impact by increasing density. These conclusions are consistent with the qualitative observations of Evans (1991) concerning the type of housing produced under the relatively tight planning regime characteristic of southern England. Also consistent are the ﬁndings of Monk et al. (1996) from four case studies, showing much higher differentials between housing and agricultural land values in areas of greater planning constraint. The present author has undertaken a series of cross-sectional studies at an inter-urban scale of analysis with an explicit planning/land supply

200

Housing Economics and Public Policy

component providing medium-term simulations of the impact of varying planning policies (Bramley 1993a & b, 1999; Bramley & Watkins 1996a). This work conﬁrms the low supply elasticities found under British conditions, and estimates that system-wide changes in land release would impact on house prices with an elasticity in the range –0.15 to –0.30. The model developed draws explicit attention to the transmission mechanism between planning policies and outcomes, part of the basis for subsequent debate with Evans (1996). As already noted, one of the central problems in empirical work in this area is that of how one actually measures planning restraint. Bramley (1998) reviews a range of measures, some objective and some subjective, and considers their logical interrelationships and some pitfalls of interpretation. There is an interesting parallel development of measures of restrictiveness in the US, reviewed and tested in Malpezzi (1996). Many of these measures are only effectively available on a cross-sectional basis, limiting the scope for time series or panel models. Yet this leads into another area of potential criticism of most of the models developed to date. By relying on cross-sections, they are vulnerable to a number of problems including the ecological fallacy, inconsistently deﬁned (heterogeneous) areas, and an inability to model dynamic effects (lags and expectations) adequately. Some US work may be better than what has been attainable in Britain, by utilising more consistently-deﬁned metropolitan housing market areas, and/or by recognising the phenomenon of varying degrees of competing jurisdictions. The problem that demandside adjustments to constraints in one area may be substantially displaced to other surrounding areas, but to differing degrees, means that econometric models applied to sets of such areas may not adequately capture price impacts. This point is picked up again below. Panel-based models with a time dimension as well as a cross-section dimension to local housing market areas provide one way forward to overcome some at least of the above limitations. Following Hsiao (1986 ch.1), it can be argued that a panel-based approach should improve the robustness of models in the following ways:

•

greatly increasing the number of observations and degrees of freedom;

•

discriminating between competing hypotheses which might be consistent with the same ‘facts’;

•

enabling some account to be taken of lags and dynamic effects, subject to the limitations of length and detail of time series; and

Planning Regulation and Housing Supply in a Market System

•

201

enabling the stripping-out of the effect of certain omitted variables (e.g. ﬁxed area effects).

In addition, one could add the following potential beneﬁts:

•

lessening the danger of results being distorted by ecological correlations and spatial autocorrelation; and

•

bringing important macroeconomic variables like interest rates into the models.

Hitherto, data availability in Britain has prevented the application of this approach. However, there is now a substantial body of secondary data on the housing market and related variables available on a consistent annual basis for local or sub-regional areas. In particular, data on land available through the planning system is now available for a period of 10 years or so in both England and Scotland. These data do not yet permit investigation of the complex transmission mechanisms between planning policies and land actually made available with planning permission, as examined in Bramley (1993b), Bramley et al. (1995) and Bramley & Watkins (1996a) using sample cross-section data. However, they enable us to model the supply of new building on land with planning permission, house prices, and certain related variables including migration. Panel-based models are beginning to appear in this ﬁeld. Mayer & Somerville (2000) provide a US example, but even here it should be noted that the measures of land use regulation are only cross-sectional. Leishman & Bramley (forthcoming) provide an early attempt at this approach using data for Scottish districts. The next sections go on to report initial tests of a similar model applied to larger zones in England.

Setting up an inter-urban panel model Recently, government-funded work has been undertaken to develop a policy model with which to predict internal migration ﬂows within England and Wales. This model, known as MIGMOD, has been developed up to an operational level. Champion et al. (1998, 2000) describe the model in detail. It comprises two main sub-models:

•

a panel model to predict gross out-migration ﬂows by age and sex for 98 zones in England and Wales, calibrated on 15 years’ data for 1983–97; and

202

•

Housing Economics and Public Policy

a spatial interaction model to distribute ﬂows across the 98 zones, calibrated cross-sectionally but separately for each origin zone and each age/ sex group.

These are combined in a user-friendly front-end to predict ﬂows conditional on assumed values of determinant variables. In order to calibrate this model, a uniquely large dataset was constructed, including a signiﬁcant number of time-varying factors for the ﬁrst stage model. This dataset is valuable for our purposes because it includes a wide range of economic, demographic, housing, social and environmental factors that potentially can inﬂuence housing market outcomes. Part of the motive for examining housing market relationships using this dataset is to explore the potential for ‘endogenising’ the housing market within the migration model. It is generally recognised that both migration and housing prices and outputs are determined simultaneously within the same system, and subject to similar inﬂuences. Actual or potential migration may be regarded as an important part of the transmission mechanism which drives the housing market and links the behaviour of different but proximate geographical areas. It would be possible to incorporate migration rates as endogenous variables within a simultaneous system designed to analyse housing supply and price responses. I do not do this here owing to technical limitations on the basic panel dataset and the complexity of the full migration model. Basically, this issue of the interaction between migration and housing is on the agenda for the next stages of research, building on MIGMOD. Nevertheless, the panel models presented do take indirect account of migration insofar as the key drivers of migration are included as explanatory variables.

Key variables For the purposes of this exercise data from one additional source has been incorporated. This is the annual ‘PS3’ return made by local planning authorities to central government, recording the stock of land with outstanding planning permission for new private housing development, measured both by the number of units capacity and the area in hectares. These data are available for the 10-year period from 1988 to 1997. There are a considerable number of missing cells at local authority district level. Rounded values were entered in missing cells by interpolation, but a count of missing cells was maintained. When aggregated to the 90 (see below) MIGMOD zones, this count was used to provide a ‘reliability’ weighting subsequently incorporated in the analysis. The main measure used in the panel dataset (PLSH)

Planning Regulation and Housing Supply in a Market System

203

is the number of housing units capacity, divided by the population in thousands and by a notional average household size of 2.5 to yield a stock measure per 1000 households (or occupied dwellings). Brief comments on several other key variables in the modelling are also in order. House prices are measured in real terms (discounting general inﬂation) and standardised using a typology of four ‘optimised’ housing types developed by the University of Newcastle. The data source is the Nationwide Building Society, one of the larger national mortgage lenders, which has maintained data over a 20-year period. New housing output is measured by private completions as per local authority returns based on building control records. These data are not wholly reliable but they are the best available. This measure (PQPR) is expressed in annual units per 100 existing dwellings (all tenures). Other economic variables available on a local and time-varying basis include: a measure of gross household income (HHINC, based on Bramley & Smart (1996) and the Ofﬁce for National Statistics (ONS) sub-national personal income series); three-year change in employment by workplace (EMPGRO); unemployment rate over working age population for males aged 30–44 (ASUNEM). Time-varying housing variables include vacancy rates (PVAC) and social housing completion rates (PQSR). National time series for mortgage interest rates (MGINTR_Z) and real GDP changes (RGDPCH_Z) are included. All time varying factors in the model are entered as levels with a one-year lag. In principle it would be possible to explore more elaborate lags. Sophisticated techniques have developed in recent years to tackle the inherent econometric problems associated with time series analysis. The relatively short length of the time dimension in our panel dataset largely precludes the more sophisticated approaches involving cointegration. However, as a precautionary measure to deal with time trends in the data we incorporate a simple time trend in the models for levels (TIME_Z). We also test the use of a model constructed wholly in ﬁrst differences. The share of new housing built on formerly urban (alias ‘brownﬁeld’ or ‘recycled’) land is available currently as a cross-sectional indicator (PNBU), although eventually this measure should be available as an annual series. A general ‘urban’ composite indicator is constructed using principal components analysis based on a measure of settlement pattern (basically, the density of existing population) and a number of other highly correlated cross-sectional indicators. The URBAN composite variable contains one time-varying component, a measure of international immigration rate, and

204

Housing Economics and Public Policy

so the differences in this variable effectively relate to immigration (hence the label IMMIGD). Other important cross-sectional variables utilised in the models include: an indicator of occupational structure weighted by gross migration propensities from the 1991 Census (OCCMIG), which broadly measures high occupational class; share of people with single marital status (SINGLE); share of non-white ethnic population (NONWH); and a composite measuring warmer and drier weather conditions (CLIMATE). London represents an extreme in many aspects of the housing market and to allow for this we also include a London dummy variable in the models, to allow for any simple shift effects not captured in other variables. A wider range of variables, mainly cross-sectional and relating to housing and environmental conditions, were tested but discarded before the ﬁnal stage of the modelling. The MIGMOD dataset includes a substantial number of ‘regional’ variables, measuring the difference between values in closely surrounding zones and the value in a particular origin zone. We retain only one of these, the measure of total population, which provides an indicator representing conurbation locations. This aspect of the model, relating to potential spatial interactions, could be developed further. The ﬁnal version of the stage 1 MIGMOD model utilised a quadratic functional form (Champion et al. (2000), but in this paper we only test a quadratic form for one variable, price, together with a couple of key interaction terms. Otherwise, the models ﬁtted are linear.

Theoretical interpretation The models reported here constitute a relatively straightforward supply and demand system. Demand is represented by an inverted demand function, with price as the dependent variable and a wide range of determinant variables including new housing supply or its key exogenous determinant, available land. Housing demand is also expected to be inﬂuenced by economic conditions (income level and growth, employment/unemployment, interest rates), environmental conditions (e.g. climate), socio-economic proﬁles (class), demographic factors, proximity to major urban agglomerations, and other sources of housing supply (social housing, vacancies). Part of the way demand variations are expressed is through migration. Supply is represented as the ﬂow supply of new completions. New ﬂow supply is modelled as a function of a smaller subset of variables, including price and available land. Bramley & Watkins (1996a & b) suggest that this function can be derived formally from models of a ‘representative’ proﬁt-maximising

Planning Regulation and Housing Supply in a Market System

205

developer holding a given stock of land and confronting given or expected price and cost levels. Since housing cannot be built without land being available with planning permission, this variable is crucial to the model, and it can be argued that the functional form of the model should reﬂect this dependence. For example, there is a case for interacting price with land available, or using a log-log formulation (we illustrate the former approach below). It is assumed that price embodies most of the demand-side inﬂuences, but there is a case for including other economic variables to capture their impact on developers’ costs. This applies particularly to interest rates but may also apply to unemployment, although in the latter case the demand effects may predominate. It is also expected that supply will be inﬂuenced by locational characteristics, particularly urban settings and land in former urban use, as these are generally believed to entail greater costs and risks for developers. Private supply may also be inﬂuenced by other sources of supply, from the social housing sector and vacancies. Clearly, price and completions are endogenous within this system. We therefore use two-stage least squares (2SLS) to estimate the structural price and supply equations, having ﬁrst regressed these on all exogenous variables retained in the model. However, we also report a comprehensive single stage OLS model for comparison. In addition, we explore some variations in functional form in relation to the key variables of price and output.

Geographical units Comment is also in order on the nature of the geographical zones used in this analysis. These comprise the 90 MIGMOD zones covering England, Wales having been discarded owing to lack of data on land with planning permission (and some other data weaknesses). These zones are the former Health Authority areas for which migration data up to 1997 were available. They are in fact simple aggregations of 366 local authority districts. In ‘shire’ areas the zones are counties; in provincial metropolitan areas they are normally single metropolitan districts (e.g. Manchester, Bolton). In London they are aggregations of 2–4 contiguous London boroughs. Previous research using local authority districts has been criticised for utilising a spatial unit that did not consistently represent a meaningful concept of local housing market (or labour market) area (see Bramley 1993a; Evans 1996). The present set-up of 90 zones is to some extent open to the same criticism, although arguably less severely. The zones are certainly not selfcontained regions, in the sense that the migration between them is not solely

206

Housing Economics and Public Policy

long-distance labour migration, but includes a varying element of housing or environment-led migration (as discussed for example in Gordon 1991). This issue has some bearing on the interpretation of the inverted demand function. The more self-contained are the zones, the more the feedback effects from land supply and new housebuilding onto market prices should be contained within them, and the less these effects leak out into surrounding areas. Given the hybrid nature of the zones, it is expected that some of the price effects of supply do leak out, and will not be fully captured in a model that treats each zone as though it were independent. Bramley (1993a, 1999) argued that, working with smaller district level units, this effect was so great that it was better to approach price impacts by conducting whole-system simulations, rather than simply reading off the coefﬁcients for district-level effects. Using the larger MIGMOD zones, it may be expected that the price model coefﬁcients will capture more of the supply impact, and the results reported below seem to bear this out. Nevertheless, these measured impacts should still be regarded as giving an incomplete picture of system-wide effects. In the longer term, it is hoped that by integrating these models with a fully interactive migration model, or by incorporating spatial interaction within these housing models, a fuller picture can be obtained. Despite their limitations, the models reported shed light on a number of policy-relevant issues, as identiﬁed earlier, including supply elasticities with respect to price and land availability, and the impact of an emphasis on urban and brownﬁeld land development.

Model results The demand side Table 11.1 presents the key regression results for house prices using the panel model set up in the manner described above. This model should as noted above be interpreted as an inverted demand function. The ﬁrst two columns (model (1)) show the fullest OLS model, incorporating both levels and differences in time-varying variables on the right-hand side (this is effectively the reduced form price equation underlying model (3)). The remaining columns give the structural demand equations resulting from the 2SLS procedure. The key difference from model (1) is that in that case ‘supply’ is represented by the exogenous land availability indicator (PLSH), while in the remaining cases this role is taken on by the endogenous new completions variable (PRPQPR) as predicted in the ﬁrst stage reduced form regression. Model (2) is a more parsimonious version including only levels indicators on

Planning Regulation and Housing Supply in a Market System

207

the right-hand side, while model (3) is a fuller version including both levels and differences for the time-varying explanatory variables. Model (4) is a model set up purely in ﬁrst differences. This has the advantage of removing possible bias associated with unmeasured ﬁxed area effects, but the disadvantage of not enabling estimation of any cross-sectional variable effects. All models in levels achieve a reasonable ﬁt to the data, with adjusted r-squared values around 0.85–0.87. This high ﬁt is characteristic of crosssectional price models in previous research. The ﬁrst differences model (4) achieves a much lower ﬁt, as expected, because it omits all cross-sectional variables and because it is more vulnerable to errors of measurement. Most of the variables retained in the model are statistically signiﬁcant. The only exceptions are the urban composite (in the 2SLS models), employment growth (in most models), some of the difference variables in model (1), and the sub-regional interaction term for neighbouring population. The urban variable might have been hypothesised to have a positive effect, reﬂecting the classic urban economic land rent model, or alternatively a negative effect indicating an aversion to urban environments; in practice neither effect is conﬁrmed. Possibly the variable SINGLE captures the central city effect. Employment growth suffers from some measurement problems associated with sampling error in the annual employment survey; it is possible that the income and unemployment variables adequately capture the underlying effects of employment. Most of the remaining variables that are statistically signiﬁcant have effects in the expected direction. House prices are strongly driven by income, higher occupational class and better climate, with negative effects from unemployment and interest rates. The only case where the sign is clearly opposite to that expected is national GDP growth, although this effect is not very strong and will be offset by the stronger income variable. The negative time trend effect is slightly surprising but again this will be offset by the income variable. The positive effect of non-white ethnic population is interesting, suggesting that the positive demand effect of ethnic population concentrations outweighs any negative effects on the demand from the indigenous population. The most interesting results in the context of our analysis of supply are those relating to the price impacts of the supply variables. In the ﬁrst OLS model, the amount of land available has a modest negative impact on house prices. In the 2SLS structural levels models, the predicted supply of new housing completions has a more substantial impact on house prices. The coefﬁcient indicates that a 10% increase in output at the mean would lower local house prices by 3.1–3.4% (see Table 11.4). This effect is substantially stronger in the ﬁrst differences model (4), 8.5% lower prices for a 10% higher output.

208

Panel model regressions for house price (90 health authority areas in England annual series 1988–97). (2) 2SLS levels

(1) OLS

(3) 2SLS levels

(4) 2SLS 1st differences

Explanatory variable

Coefﬁcient

t-stat

Coefﬁcient

t-stat

Coefﬁcient

t-stat

Coefﬁcient

t-stat

Constant *URBAN *IMMIGD *HHINC

122176 6067 15144 107.9 3.66 –1551 832 –128.0 2084 1897 –2008 –684 –64.4 43.5

7.60 3.91 1.46 7.74 0.05 –5.85 1.23 –0.54 1.95 1.87 –4.10 –0.90 –2.60 0.68

126857 –1247

8.68 –0.83

–4.76

–15621

–1.50

10.55

50.0

0.71

–2471

–9.89

–43.1

–0.19

9.48 –0.67 1.22 13.68 0.60 –8.37 1.07 –0.61

–6074

147.0

159798 –992 12836 155.4 41.9 –2339 727 –145.1

–6299 –784

–9.67 –3.78

–28906

–10.29

–31492

–11.51

–78655

–6.70

*ASUNEM *EMPGRO *PQSR *PVAC *PLSH *PRPQPR

Composite dense urban Difference in urban Household income level Difference Unemployment level Difference Job growth % change Social hsg completions level Difference Vacancy rate level Difference Land available with OPP1 level Difference Predicted new completions level Difference

Housing Economics and Public Policy

Table 11.1

RGDPCH_Z TIME_Z LONDONDU Adj r-squared F N HPRICE

Urban land share Occupational class index Single marital status Nonwhite ethnic population Composite warmer/drier Neighbouring popn level Interest rate level Difference Real GDP growth % Time trend London dummy

205.8 25940 339 131.2 2389 182.4 –5201 4603 –2071 –7894 –3388

Dep. variable mean value

0.866 194.9 719 76930

8.37 6.13 1.22 1.55 3.66 0.26 –7.49 5.42 –2.60 –14.69 –1.61

18241 1529 184.5 5092 974 –4128

4.19 5.66 2.16 8.18 1.36 –10.56

–563 –7861 –9118

–1.89 –23.44 –3.90

0.853 298.8 719 76930

17722 1430 146.3 5028 947 –5920 5381 –1611 9042 9755 0.861 248.2 719 76930

4.08 5.19 1.73 7.75 1.36 –8.78 6.47 –2.04 –16.86 –4.26

–2531

–5.65

0.227 31.2 719 875

Note: * indicates local time-varying factors; all level variables with 1 year lag; _Z indicates national time series; all regressions weighted by population and reliability index. 1

OPP = outline planning permission.

Planning Regulation and Housing Supply in a Market System

PNBU OCCMIG SINGLE NONWH CLIMATE TOPOPN MGINTR_Z

209

210

Housing Economics and Public Policy

It is difﬁcult to compare this result directly with previous studies, because of differences in exactly what is being measured on the supply side. The elasticity of price feedback with respect to output of –0.33 above is well above the –0.041 for secondhand sales and –0.191 for new sales found in Bramley (1993a & b). It is rather more similar to the result yielded by cross-sectional price regression equations at district level in the second study by Bramley & Watkins (1996a ch.6). It is also rather stronger than the results obtained by Leishman & Bramley (forthcoming) applying a panel 2SLS model to Scottish districts, which yielded an elasticity of house prices with respect to predicted completions of –0.225. The signiﬁcance of the size of zones for this measure of feedback was discussed above, where it was suggested that the larger MIGMOD zones would be expected to capture a greater degree of feedback, and these ﬁndings appear consistent with that view.

The supply model We can now turn to the supply model results that are summarised in Table 11.2. These are of greatest interest in the context of this chapter. It should be remembered that supply here is deﬁned as the annual ﬂow of new private housing units. As with house prices, we compare an OLS formulation (model (1)) with several 2SLS equations. There is some evidence to support a model involving an interaction between price and land availability, together with a quadratic term in price, so we include both this version (2B) and a simpler linear form (2A) of the structural equation. The interaction of price (or more strictly, gross development proﬁt) with land availability featured in the previous work by this author (Bramley 1993a & b; Bramley & Watkins 1996a & b). The rationale is simply that houses cannot be built without land with planning permission, so that the effect of price may be expected to work in conjunction with land supply rather than independently. The data appear to support this formulation in the new panel model (comparing (2B) with (2A)), as it did in the previous cross-sectional models. The overall ﬁt of the supply models is poorer than that of the price models, with about half the variance in output levels being explained. This is consistent with all previous comparable work. Supply is more difﬁcult to model, perhaps because there are more unique local factors making for a lumpiness in annual output rates not related to the broad, systematic variables included in the model. There are also some grounds for doubt about the accuracy of the recording of the timing of housing completions, imparting greater random error. Mention has already been made of the weaknesses in the measurement

Planning Regulation and Housing Supply in a Market System

211

of the key driver of land supply, and of the reliability weighting incorporated in the regressions. The ﬁt is particularly poor for the ﬁrst differences model (4), although some of the key variables remain signiﬁcant. The model results broadly do show that both price and land availability have a positive impact on the level of housebuilding output. However, price is negative and not signiﬁcant in the OLS model (1), while land supply is negative and insigniﬁcant in the ﬁrst differences model (4). The ﬁrst ﬁnding probably reﬂects the inappropriateness of OLS when price is clearly endogenous and strongly related to a number of other variables included in this model. The second ﬁnding may reﬂect data weaknesses in the context of a poorly ﬁtting model. It may also reﬂect a spurious negative correlation effect, because increased completions would, for any given level of new planning permissions, tend to deplete the existing stock of completions more quickly. Looking at the 2SLS models in levels, when price and land are entered separately as linear variables, as in models (2A) and (3), both are signiﬁcant and positive, but the land effect is much more robust in its signiﬁcance. When entered as an interaction term (model (2B)), this improves the overall model ﬁt and the combined variable is positive and robust in terms of its signiﬁcance. The squared price terms appear to have a negative effect, but this is not statistically signiﬁcant. Thus, there is some support, but not robust evidence, for the argument that the price-supply function weakens or bends backwards at higher price levels (as discussed in Pryce 1999). It is interesting to note that in model (2B) the separate land term has a negative effect, although this is of marginal statistical signiﬁcance. This suggests that allocating more land in a situation where prices are low could have a low or even negative impact on output. Model (3) conﬁrms the broad magnitude and signiﬁcance of both price and land effects when a wider range of explanatory variables, including difference terms, are included. It is noteworthy that both the level and change in land supply have positive effects of a similar order of magnitude. What are the implications of these ﬁndings for housing supply elasticities? Calculating values at the mean and allowing for the net effects with any interactions, the supply elasticities revealed by this model are shown in Table 11.3. The price elasticities from the 2SLS levels models lie in the range 0.36–0.38. The ﬁrst differences model (4) yields a higher ﬁgure of 0.585, although the shortcomings of this model should be remembered. The elasticities of supply with respect to land available with planning permission lie in the range 0.25–0.31, except in the case of the ﬁrst differences model where it is as noted negative and insigniﬁcant.

212

(1) OLS

(2A) 2SLS

(2B)2SLS interaction

(3) 2SLS levels

(4) 2SLS 1st differences

Explanatory variable

Coefﬁcient t-stat

Coefﬁcient t-stat

Coefﬁcient t-stat

Coefﬁcient t-stat

Coefﬁcient

t-stat

Constant *HPRICE

1.873 –1.12E–6 –1.70E–6

1.293

0.271 4.16E–6

–3.74E–2 6.34E–6

–4.35 2.97

House price level Difference *HPRICE_Q House price squared *HPRICE x PLSH Price x land interaction *PLSH Land available OPP1 level Difference *PQSR New social hsg comps level Difference *URBAN Composite dense urban *IMMIG Difference *HHINC Household income level Difference *EMPGRO Job growth % change *ASUNEM Unemployment level Difference

6.88E–3 5.71E–3 2.29E–2 2.64E–2 –7.91E–2 5.31E–2 2.14E–3 2.51E–4 9.52E–4 –1.23E–2 –1.45E–2

4.16 0.444 –1.14 3.94E–6 –1.25

2.03 5.42

–2.31E–7 1.20E–7 10.63 –2.38E–3

10.97 5.78E–3 3.49 0.85 –7.34E–2 –3.23 –5.50E–2 1.02 –2.00 –9.87E–2 –4.08 –8.76E–2 0.20 5.83 0.14 0.16 –1.78 –5.79E–3 –0.84

5.99

–0.53 6.02 –1.59 6.56E–3 5.15E–3 –2.47 –2.08E–2 4.48E–4 –2.96 –0.104

0.66 2.99

12.23 3.18 –1.13E–3 –0.75 0.02 7.83E–2 –3.44

–1.15 5.91E–4 0.10 –1.14E–2 –0.72

–0.86 3.54

Housing Economics and Public Policy

Table 11.2 Panel model regressions for new private housing completions (percent of total housing stock; 90 health authority areas in England annual series 1988–97).

*PVAC

TIME_Z LONDONDU Adj r-squared F N PQPR

–1.13E–2 4.28E–3 –8.16E–4 –0.404 1.93E–2 –1.38E–3 1.76E–2 –6.54E–3 –5.09E–2 6.15E–2 –5.08E–2 –0.20

Dep. variable mean value

0.519 30.8 719 0.8756

–0.91 0.22 –1.25 –3.69 2.75 –0.65 1.05 –0.37 –2.53 2.70 –3.06 –3.76

1.04E–2

0.81

8.60E–3

–2.54E–3 –4.08 –2.31E–3

–8.54E–3 –4.11E–3 –0.46 –3.37E–2 9.33E–4 –0.154 0.439 63.6 719 0.8756

0.08 –2.03E–2 –3.28 –4.77E–2 0.511 63.5 719 0.8756

0.70

4.61E–3 0.34 6.53E–3 0.33 6.96E–3 –3.36 –1.93E–3 –2.69

0.46

–0.50 –1.89E–2 –1.08 –5.04 –4.26E–3 –0.30 2.81E–2 1.92 –3.18E–2 –2.24 1.22E–2 0.74 –1.01 –0.129 –2.83

–5.62

0.498 45.6 719 0.8756

0.049 8.45 719

Note: * indicates local time-varying factors; levels variables with 1 year lag; _Z indicates national time series; all regressions weighted by population and reliability index; price terms are based on predicted price from reduced form in 2SLS models. 1

OPP = outline planning permission.

Planning Regulation and Housing Supply in a Market System

PNBU OCCMIG SINGLE NONWH CLIMATE TOPOPN_Y MGINTR_Z

Vacancy rate level Difference Share of former urban land Occupational class index Single marital status Nonwhite ethnic population Composite warmer/drier Neighbouring population Interest rate level Difference Time trend London dummy

213

214

Housing Economics and Public Policy

Table 11.3 Elasticities of supply (private completions) from different models (evaluated at mean). Model

Price elasticity of supply

Land elasticity of supply

Urban land elasticity of supply

(1) OLS (2A) 2SLS simple (2B) 2SLS interactive (3) 2SLS levels and differences (4) 2SLS 1st differences

–0.102 0.362 0.365 0.381 0.585

0.300 0.252 0.308 0.286 –0.049

–0.062 –0.194 –0.149 –0.147 n/a

It is clear that the main message of this new model is that housing supply elasticities in England in this period are relatively low. For example, the estimates in Table 11.3 are well below the results reported in Bramley (1993a) which were of the order of 0.8–1.1, based on cross-sectional analysis at district level for 1988. However, this new low estimate is more comparable to recent national and regional results from Meen (1998), and also to other cross-sectional estimates (unpublished) for the 1990s from this author, and the Scottish results from Leishman & Bramley (forthcoming). These elasticities are markedly lower than those reported in US literature. Low supply elasticities imply that demand shocks are more likely to generate larger price ﬂuctuations rather than output responses. They also imply that it will be difﬁcult to achieve large increases in owner occupation (Meen 1998). A further implication is that any demand-side subsidies would tend to be capitalised into house prices (this may be of contemporary signiﬁcance in the context of policies to subsidise access for key workers). One hypothesis that may go some way to explaining the apparently falling supply elasticities might be that there are increasingly important variations on the cost side of the development equation, upon which we do not have systematic data. The brownﬁeld land variable (PNBU) attempts to pick up one aspect of this. But there is an increasingly signiﬁcant phenomenon of planning agreements being used to secure the provision and ﬁnancing of large elements of infrastructure, and also other costly elements such as ‘affordable housing’. These onerous requirements are most prominent on large developments on greenﬁeld sites in the south of England. This means that the relationship between price and ‘gross development proﬁt’ becomes increasingly tenuous. Pending the collection of systematic data on planning agreement costs, this remains speculation. The land supply elasticities are also surprisingly low. This may reﬂect in part the measurement inadequacies mentioned above. However, more substantively, it indicates that just allocating extra land for housing and giving it

Planning Regulation and Housing Supply in a Market System

215

planning permission is not sufﬁcient to ensure that it will all be built upon in the short to medium term. The lower values for this parameter than in previous research (typically around a half) may be in part a reﬂection of the larger geographical units in this study; the ‘steering’ effect of land allocations may be stronger at the lower district scale than at the broader MIGMOD zone scale. It is possible to combine these estimates of land supply elasticity of output with the estimates of supply-price feedback derived from the price model, as shown in Table 11.4. The implied price impact of a 10% increase in land availability would be of the order of –0.8% to –1.0%. This elasticity of price with respect to land supply of around –0.1 is rather below the range of estimates (–0.15 to –0.30) presented in Bramley (1999), but the latter were based on whole system simulations of general increases/reductions rather than single-zone scenarios. A number of other variables in the supply model are also of policy relevance. The share of former urban land (brownﬁeld land) within the supply (PNBU) has a consistently negative effect on output. Table 11.3 shows that the elasticities are of the order of –0.15 to –0.19. This is consistent with the widespread assumption that brownﬁeld sites are likely to be more costly, difﬁcult or risky to develop. The government is seeking to increase the share of brownﬁeld land in housing provision; this evidence suggests that this policy will carry some cost in terms of reduced output from a given land stock, with some feedback effect on to house prices. For example a 10% increase in the urban land share would reduce output by 1.49% at the mean. Combining this elasticity of –0.149 with the –0.33 elasticity of price with respect to output would give a price impact of +0.49%. Another way of looking at the urban land share effect is to regard it as a cost mark-up. The 10% rise in urban land share has the same impact on output as a 3.91% fall in prices (model (3)), equivalent to £2955 for an average housing unit. This could be a measure of the marginal cost effect of brownﬁeld land, although as such it looks rather on the high side. Table 11.4

Elasticities of local price feedback (evaluated at mean).

Model

Supply-price feedback

Impact of 10% more land

(1) OLS (2A) 2SLS simple (2B) 2SLS interactive (3) 2SLS levels & differences (4) 2SLS 1st differences

–0.309 –0.309 –0.337 –0.849

–0.30% –0.78% –0.95% –0.96% –0.59%*

* derived from reduced form.

216

Housing Economics and Public Policy

Also consistent with this evidence is the signiﬁcant negative coefﬁcient on the URBAN composite variable; output is lower, other things being equal, in more urban settings. The London dummy effect is also negative, which can be seen as part of the same general story. Lastly, the time trend variable is signiﬁcantly negative in all the models. This can perhaps be related to the general policy concern being voiced by the housebuilding industry at the moment that the overall level of housebuilding output in England is at record low levels in the late 1990s, despite the buoyant economic climate and high or rising prices.

Conclusions The supply side of housing has been relatively neglected in housing economics, despite presenting a range of important analytical and policy challenges. Neither of the dominant traditions in econometric work has got fully to grips with the key supply constraints associated with land use regulation. Ways forward for research may involve more micromodelling of supply, more incorporation of insights and measures from qualitative/institutional work, more incorporation of spatial interaction perspectives, and the combining of time and cross-sectional dimensions. This chapter exempliﬁes the latter approach by exploring the application of a panel model to analyse the impact of the amount and type of available land on new housing supply and market price outcomes in England. Planning regulation has survived the wave of deregulation in Britain since 1980, while in the US there is a pronounced move towards more interventionist urban growth controls. The policy issues to which these interventions respond – environmental, ﬁscal, social and urban – are sufﬁciently important that simple calls for getting rid of all regulation are unlikely to be acted upon. What are needed are more intelligent, well-informed and well-implemented policies, and urban economics has a role in providing the information and intelligence required. Economists agree about the general nature of the impact of planning regulation on housing supply and markets. What is unclear is the relative magnitude of different effects, some of which are potentially contradictory. Existing empirical work provides some indications of the likely effects, but the magnitude and interpretation of these results varies. Thus there is a need for better-speciﬁed models.

Planning Regulation and Housing Supply in a Market System

217

Results of the initial application of the inter-urban panel model in England support a number of propositions. Firstly, the price elasticity of new build supply is low in this country, and may be getting lower. Secondly, land availability through planning is a strong driver (or steerer) of development. Thirdly, greater supply does reduce house prices, although the local impact of this is quite moderate. Fourthly, increasing the emphasis on urban and brownﬁeld land does carry a signiﬁcant output penalty, with a modest consequent impact on house prices.

Acknowledgements This chapter draws heavily on data generated within a research project commissioned by the Department of Transport, Local Government and the Regions (DTLR, formerly DETR) concerned with modelling migration ﬂows in England. The author was a member of this research team but would like to acknowledge the permission given by DTLR for use of these data, as well as the help and encouragement of other members of the research team, particularly Tony Champion, Stewart Fotheringham and Stamatis Kalogirou of the University of Newcastle, and other colleagues at Universities of Newcastle and Leeds. Particular thanks are due to the Nationwide Building Society for permission to use house price data derived from their lending records. The author would also like to acknowledge support and encouragement provided by Duncan Maclennan for earlier efforts in this area. Responsibility for the conclusions and interpretations drawn, as well as any errors or omissions, remain the responsibility of the author.

12 Economics and Housing Planning Tony O’Sullivan1

Introduction There has long been a general presumption in the UK that local authorities have a role in the planning of housing provision.2 This role encompasses all tenures, including housing directly provided by local authorities, social housing provided through non-proﬁt making organisations, and housing supplied solely through private markets for rent or to own. This state of affairs is in marked contrast to the situation in the USA where such a role for municipal government is not considered at all feasible or appropriate. Maclennan (1991) provides a succinct case for the need for local strategic housing planning – although, as he notes, this does not necessarily imply municipal government should have the responsibility for undertaking it. The case is essentially that:

•

in efﬁciency terms, local housing markets are so pervasively beset with problems of market failure, arising from long product life, slow supply adjustment, information imperfections and external effects, that markets operating without a socially determined ‘game plan’ to orient them are bound to be sub-optimal; and

•

in equity terms, the product ‘housing’ looms so large in the welfare of all households that the consequences of local housing markets getting it wrong are socially (not to mention politically) unacceptable.

Almost by default it has come to be seen as a local authority responsibility in the UK. Among other things, this has involved an implicit assumption that local authorities are actually capable of doing the sort of planning en-

Economics and Housing Planning

219

visaged. Taking their responsibility at face value, local authorities are as a consequence confronted with a demanding set of analytical tasks. Effectively, local authorities are expected to make a credible forecast of demand for housing for their areas of jurisdiction over a 3–5 year planning horizon. This forecast should nest within a 10–15 year forecast of the overall infrastructure and land requirements (commercial and industrial as well as residential) of the regional economy in which an authority operates. By making allowance for future trends in income growth and distribution, consumer preferences and market supply adjustments (the latter heavily inﬂuenced by land use planning), authorities are expected to establish the required volume and location of subsidised social renting accommodation. At the same time, they must monitor physical and social obsolescence of existing housing, and encourage, fund or take (where they own the housing) appropriate offsetting action. In principle, to deliver against these expectations involves the application of a considerable amount of economics in local housing contexts, covering (among other things) articulation of new and second-hand market structure and evolution, preference measurement, demand forecasting and tenure choice estimation. In this chapter I consider how well local authorities dispense the housing planning function. The next section describes how this responsibility on authorities evolved over the greater part of the twentieth century, and the current position. Thereafter, I present evidence that demonstrates a gulf between the requirements and reality of local authority housing planning. This conclusion is reinforced by consideration of the contribution made by land use planning (also a local authority responsibility) to housing plans. The penultimate section considers reasons for the gap between requirement and reality. It looks at the economic toolkit implicitly required to effect housing planning and compares this with available economic tools. It also considers the behavioural incentives and disincentives that face local authorities in this area. The evidence available raises questions concerning the application of the rational comprehensive planning model to housing. The ﬁnal section offers some concluding thoughts.

The policy and institutional framework The emphasis of housing policy in the UK has changed markedly over time as a consequence of both change in perceptions of the nature of housing problems, and the actual and perceived success or failure of speciﬁc policy initiatives. (Galster & Daniell 1996, show this has been true for the US

220

Housing Economics and Public Policy

also, and demonstrate the role social science has played there in a housing planning/policy making context.) In general, one consequence of this evolution has been that planning expectations on UK local authorities have steadily increased over time.3 In the years immediately preceding World War II, local authorities had two principal sets of responsibility: provision and management of houses to let, and improvement of private sector stock condition. The latter included slum clearance and grant-aided improvement, both of individual dwellings and area improvement schemes. Severe shortages of housing developed during the war years. For the two decades following the war there was consequently a need for more house building virtually everywhere. Over this period ‘catching up with shortages took much longer than expected because the number of households increased faster than anyone had foreseen. Where to build was a matter of equity between people living in different parts of an authority’s area rather than anything to do with the economics of housing markets’.4 Shortages eased after the mid-1960s. The 1970s then saw the emergence of the idea of a ‘comprehensive housing service covering potential as well as existing consumers, with local authorities taking the broadest view of their statutory housing functions’. (Smith 1989, p. 61) An early signal from central government that local authority focus should broaden and deepen from a singular emphasis on building as many ‘social renting’ units as possible came from Scotland. The Scottish Housing Advisory Committee published a report in 1972 on the need for a comprehensive approach to area housing needs assessment (Scottish Housing Advisory Committee 1972). This body noted: ‘The approach we are recommending goes well beyond the assessment of the need for local authority housing. In our view, the responsibilities of local authorities should extend far beyond providing for the needs of those who are actually housed by them: they should be looking at the totality of needs – hidden needs, needs which are not being met, and needs which may arise in the future… This includes needs of those who do not wish or cannot be housed in the public sector. The provision of a council house is one way of meeting a need; it is not the only way. The implication is that

Economics and Housing Planning

221

local authorities should make an assessment of the need for other types of housing provision. In practical terms this involves, in particular, ensuring that sufﬁcient suitable land is allocated for private building …’ (Scottish Housing Advisory Committee 1972, pp. 7, 12) In England, central government introduced a new system of local authority ‘housing investment programmes’ from 1978/9 onwards. This framework quickly became linked to the idea of a strategic housing responsibility for local authorities that was to be based on comprehensive local housing plans (Department of the Environment 1977; Smith 1989). Impressive ofﬁcial guidance on how to do the required analysis subsequently emerged (Department of the Environment 1980), and was not updated until quite recently (Bramley et al. 2000). In Scotland, the 1977 Green Paper on Scottish Housing Policy (Scottish Ofﬁce 1977) proposed that housing plans prepared by Scottish local authorities would form a continuing, informed database on Scottish housing. Although it would be 2001 before such plans (or ‘local housing strategies’) became a statutory responsibility on authorities in Scotland, from 1977 housing capital allocations to authorities were to be based on locally-prepared and detailed triennial plans covering all sectors (Maclennan 1989). This was supported by the publication in 1977 of detailed guidance to assist ofﬁcials of local authorities engaged in preparing comprehensive assessments of housing needs in their areas (Scottish Development Department 1977). Guidance on how to analyse local housing markets has not subsequently been ofﬁcially updated, although in the early 1990s Scottish Homes5 developed an approach jointly with Glasgow University, and subsequently published it (Maclennan et al. 1998a). More generally, there has been a continuing stream of commentary and contribution from academic and other sources (see, for example, Audit Commission 1992; Van Zilj 1993; Maclennan 1991; Fordham & Brook 1995; DETR 1997; Hawtin 1996; Wood & Preston 1997; Goss & Blackaby 1998). In the 1980s council house building was cut back to virtually nothing. New social provision was henceforth to be predominantly made through housing associations –voluntary sector institutions that develop housing through combining public sector capital grant with private loans. The reduced provider role for local authorities was rationalised as part of a further shift by local authorities (under government pressure) into a ‘strategic’ housing role. Subsequent developments in policy such as encouragement to transfer remaining authority-owned stock to other landlords and to secure social housing from planning gain arising from private residential

222

Housing Economics and Public Policy

development have further increased the pressure on authorities to develop this strategic role. These policy developments have been accompanied by a changing operating context that has in the last decade been increasingly complicating the task. Market provision now accounts for around four-ﬁfths of all housing in the UK. Moreover, from a general position of national and local shortage, the housing situation has developed into one where there is a complex and shifting pattern of areas of high and low demand for housing relative to supply. The current position is that there is a statutory responsibility on UK local authorities, within a framework of policy set by central government, to prepare and submit housing plans, or ‘local housing strategies’ with a 3–5 year timeframe, on a regular basis. Moreover (Bramley et al. 2000), there is general agreement among practitioners, experts and published guidance that the main and essential elements of a local housing strategy are:

•

a broad vision for the area covered by a local authority and its housing;

•

consultation with local ‘stakeholders’ including private sector organisations on trends, problems, issues and appropriate ways of addressing them;

•

sound, up-to-date information on the local housing system; full tenure coverage and understanding of its key dimensions, including interactions between tenures;

•

clear priorities for tackling speciﬁc problems that have been identiﬁed through analysis, and clear, appropriate objectives based on these priorities;

•

development and appraisal of options to achieve these objectives; and

•

a clear implementation framework for the selected options.

Housing plans in practice Two studies that span the 1990s experience of local authority housing plan production allow us to investigate the analytical quality of housing plans produced in the UK. The ﬁrst was a study by Varady (1996) that has the advantage of giving a view from totally outside the British framework.6 Varady took a case study approach focused on four British cities: Glasgow and Dun-

Economics and Housing Planning

223

dee in Scotland, and Birmingham and York in England. The housing plans of these four cities, identiﬁed as best practice examples on the basis of ofﬁcial and academic advice, were variously produced between 1990 and 1993. Varady evaluated the plans against seven criteria:

•

quality of ‘mission statements’;

•

extent to which plans got beyond a focus on purely low-income housing;

•

approach(es) to measuring housing problems;

•

methods of establishing priorities;

•

extent of private sector and consumer consultation;

•

coverage of implementation issues (e.g. co-ordination between housing and planning agencies); and

•

degree to which neighbourhood geography ﬁgured in the plans.

For all that Varady commends the British system relative to the situation in America, he produced some rather trenchant criticisms of what were, after all, best practice examples. On the scope of these plans, he noted that all four plans discussed shelter (lowincome housing) needs, but were less than adequate in addressing the middle-income section of the market (a criticism still being made with respect to Glasgow by Robertson (1998), and with considerably more venom). With regard to analytical content, Varady used a housing plan for Dundee produced by Grant et al. (1976), a plan produced with much central government support, and the basis for the guidance produced in 1977, as a benchmark. He noted that none of the plans considered achieved the analytical comprehensiveness of the 1976 Dundee plan, and that for all four there was in fact little connection between housing market analysis and the remainder of the plan. With respect to the establishment of priorities and the selection of options, in none was there any indication whatsoever of how the priorities reﬂected in the plans were developed. Varady noted that all four plans discussed at length what policies each authority intended to implement but both the Scottish city housing plans suffered from lack of co-ordination with development (land use) planning. On issues of neighbourhood, or the internal structure of urban markets,

224

Housing Economics and Public Policy

performance was again patchy. Only the Birmingham plan paid any real attention to geographic housing issues. Glasgow and York had little in the way of any neighbourhood focus, while it was noted that ‘the Dundee Housing Plan is not a geographically oriented document’ (Varady 1996, p. 281) How had the situation changed by the end of the 1990s? For the answer to this we turn to a study by Blackaby (2000), which provides an evaluation of the extent to which local authority housing assessments in England and Scotland, conducted at the end of the 1990s, cover private sector housing issues.7 In this study, Blackaby considered the work done by or on behalf of 34 local authorities that responded to a more general invitation to authorities to participate. Blackaby (2000) asserts local housing strategies should include:

•

coverage of the private housing market;

•

attention to interaction between tenures;

•

consideration of the scope for the private market to meet housing need;

•

identiﬁcation of potential improvements to the efﬁciency and effectiveness of the private housing market; and

•

evidence of involvement of private agencies in the housing requirements assessment process.

Blackaby found that relatively few of the assessments contained information on the private market. Where present, discussion was superﬁcial. None of the assessments contained any analysis of the characteristics of the key players in the local housing market. Many included some price and rent statistics, but these were generally restricted to single point in time averages. Further, Blackaby concluded little or no attention had been given to questions of tenure interaction. With respect to the potential role of the private market in meeting housing need, he noted analysis might be expected to cover the extent to which households require subsidy and in what degree, the extent of households in mortgage difﬁculties, and the tightness of the private rental system. Again, limited coverage of these topics was evident.

Economics and Housing Planning

225

In terms of analysis of the structure and operation of the local housing market from an efﬁciency point of view, Blackaby found again that this was largely absent. Indeed, in most cases local authorities had not even seen private housing organisations as relevant sources of intelligence in constructing local assessments. Examples of good practice were very much the exception to the rule, with only one assessment attempting a systematic review of the current state of the housing market, and only one attempting a detailed analysis of supply/demand balances across all tenures in any detail. Echoing the ﬁndings of Varady in another way, Blackaby also found little evidence of concrete policy conclusions arising from the assessments undertaken, or of any actual policy adjustments made on the basis of the assessments. The ﬁnal conclusion drawn from the study was that: ‘Most reports included in this review fall short of the comprehensive approach to assessment envisaged by much of the ofﬁcial guidance … Only a minority of documents sought to draw out the relationship between demand and supply across all tenures or to discuss in any depth the current state of the housing market.’ (Blackaby 2000, p. 34)

The land use planning system and housing planning An important complementary process to housing planning in the UK is the land use ‘planning system’. This is also a responsibility of local government. Since 1947, local authorities have been obliged to prepare plans for the use and development of land. The context for and priority accorded to the planning system have changed markedly since its introduction, but the basic legislative principles underlying the operation of the system have remained unchanged (Davies 1998). There is not enough space here for a full discussion of the planning system and its relationship with housing planning. However, it is important to note that, historically, the most sophisticated analysis of various dimensions of local housing markets have actually tended to arise through the planning system per se rather than through plans produced by housing departments. Household number forecasting, housing market boundary identiﬁcation and tenure choice issues are often addressed by local authority land use planners rather than by their housing planning colleagues. Even so it is widely felt that the economic content of land use planning is very limited (Cullingworth 1997b).

226

Housing Economics and Public Policy

Brieﬂy, it needs to be noted here that for the most part future household numbers are established on the basis of demographic projection, with few attempts to allow for economic inﬂuences (and none that would be considered successful).8 In terms of the geographic boundaries of private housing markets, it has been argued that land use planning policies should not only reﬂect the functional boundaries of local housing markets, but also the existence of any underlying sub-market structures (Jones & Watkins 1999). However, only the most sophisticated of Structure Plans attempt to broach the issue, even in a cursory fashion (School of Planning and Housing 2001; Jones 2002). Maclennan & Tu (1996) note the paucity of understanding of the structure and functioning of major UK housing markets.9 Lastly, with respect to estimating current and future market demand, the School of Planning and Housing (2001) recently concluded that little attention has ever been paid by planners in Scotland to house prices as a source of evidence – a conclusion that is borne out in an English context by the work of Monk (2000b) and Monk & Whitehead (2000).10

The roots of the problem Overall, then, it would appear that the conclusion drawn by Maclennan et al. (1994), to the effect that housing plans lack content, method and credibility, still holds true. In particular, housing plans or local housing strategies are silent or superﬁcial in their treatment of private housing issues, and in their use of economic analysis to understand past trends, present circumstances or future prospects. If after so many years the analytical basis of housing planning remains so patently underdeveloped, it is reasonable to ask why. The approach to housing strategy development taken and promoted in the UK is one of comprehensive rational planning (More 2002). This approach emphasises extensive use of analytical tools, and promotes the role of forecasting as a basis for plan development. Accepting the validity of this approach to housing planning for the moment, one can note that practical analytical tools to assist housing planners largely remain to be developed. Addressing the matter explicitly in terms of time and space, Table 12.1 suggests some questions where economics might be expected to furnish answers. In other words, therefore, those charged with the development of local housing strategies might look to economists to:

•

provide methods for isolating the functional boundaries of local housing markets at a point in time and for tracking the way in which they change over time. This would involve accounting for the key determinants of

Economics and Housing Planning

227

Table 12.1 Economic aspects of housing markets. Present conditions

Future prospects

Housing market identiﬁcation

A What are the current boundaries of the local housing market and what factors determine them?

B How are these boundaries adjusting over time and what factors are driving this? Are there cycles and/or trends evident within the local housing system? What is the trajectory of local demand, given regional and national employment trends?

Internal system structure

C What is the structure of housing supply and demand within the system? How do the various housing sectors interact at the margin? What is the spatial structure of supply and demand across the various sector price distributions? What is the nature of the local supply system in terms of capacity and ability to read and respond to demand signals? Do subsystems or sub-markets exist and if so why? What problems are failure of the market and non-market sectors to equilibrate causing?

D Are overall trends likely to generate speciﬁc rubbing points at the margins between sectors, or in speciﬁc localities? Is the supply base within the system adequate to emergent demand conditions?

these boundaries (the structure and evolution of local labour markets) and the impact on them of signiﬁcant policy positions taken by local authorities (particularly with respect to the spatial structure of planning permissions);

•

construct models of household formation and housing demand within given geographies that are also underpinned by satisfactory representations of labour markets (effectively building in sensible approaches to migration forecasting), and which allow for search, income and price dynamics (Maclennan 1986a);

•

map the structure of the housing industry across and within local housing markets, providing meaningful representations of likely decisions regarding supply from new and existing supply sources under different market conditions and prospects;

•

develop typologies of demand and supply that permit the reduction of the overall complexity of a ‘local housing market’ into meaningful but more homogeneous groupings of spatially referenced demand and supply;

228

Housing Economics and Public Policy

•

establish methods for identifying, tracking and forecasting spatially referenced sub-market emergence and resolution (this does not simply refer to private market housing, but encompasses the patterns of behaviour that occur in the social rented sector);

•

develop methods for identifying local market failures, their underlying causes and appropriate solutions;

•

create policy simulation models that using a credible behavioural base allow the examination of ‘what-if’ scenarios that support deeper understanding of the economic consequences of speciﬁc policy positions; and

•

provide techniques for evaluating alternative options for achieving speciﬁc objectives, and more generally for setting priorities and allocating resources across programmes of expenditure.

Against this list, one can note ﬁrstly (Table 12.1 A) that while there are sophisticated conceptual representations of the scale of urban housing markets (DiPasquale & Wheaton 1996) these do not translate into ready frameworks for practical application in speciﬁc housing planning contexts. One searches in vain in the literature for applied studies on the overall scale and functional boundaries of local housing markets in the UK (Jones 2002). As a consequence, while most accept that functional housing market areas are more appropriate than formal administrative areas for assessing the scale of housing demand and supply imbalances, we cannot satisfactorily identify them. Nor do there exist compelling forecasting models of housing demand based on functional geographies. More broadly (Table 12.1 B), there is a lack of any explicit dynamic analysis of local housing markets that would allow housing planners to separate function from dysfunction in observing their operation. Long-run trends, steady-state operation of housing markets as individual households move through their life cycles and the impacts of one-off shocks are all likely to manifest in superﬁcially similar behaviour,11 but the tools to ﬁlter the evidence remain to be developed. Practical analysis of the internal structure of local housing markets (Table 12.1 C) is hardly more ﬁrmly founded, although there is more of it around (Jones et al. 2001b). To understand the operation of a local housing market one must be able to decompose it into more substitutable from less substitutable housing ‘products’. One must also be able to establish whether speciﬁc sub-markets exist, and where they do, to map their spatial structure. Efforts to characterise the internal structure of local housing markets include work by Maclennan attempting to deﬁne ‘product groups’ based on the attribute structure of existing stock within Scottish cities (see Maclen-

Economics and Housing Planning

229

nan et al. 1994). Most UK studies attempt to identify spatial sub-markets using a hedonic approach, which raises considerable well-known conceptual problems.12 There is little in the way of analysis of household behavioural responses to administered systems of housing supply. There is also a dearth of understanding of market housing supply issues, from the question of how best to characterise regional, local market or intra-market supply structures, to the development of meaningful representations of the behaviour of supply agents under different market conditions. Finally, there is the question of intra-market dynamics (Table 12.1 D). There has been a substantial US-based model-building effort since the 1970s (particularly at the National Bureau of Economic Research, and the Urban Institute – see Rothenberg et al. (1991) for a review of these models). Work following the lead of these contributions in a UK context includes that of Ballas & Clarke (1999), and the demand forecasting work of Gibb et al. (2000). Such models can be used to simulate the allocation of households to housing stock. The Gibb et al. study forecasts demand for both social renting and owner occupation in Glasgow City to 2009, based on a disaggregation of demand into three areas, and as it breaks down between net migration into the city, new household-based demand, and moves between the three areas by tenure and house type (see Chapter 2 for a fuller discussion). There is no doubting the value of such work, or that key issues can be addressed using it. However, the geographies for the work are still dominated by formal areas rather than functional; the explanatory variables used remain constrained by what is available rather than what is relevant, and the quality of forecasts (where they exist) for some key explanatory variables remains extremely low. The foregoing does not add up to a complete analytical toolkit for the housing planner concerned to get a defensible handle on the state of affairs in a local housing market, existing housing needs, or immediate prospects for the housing system. If, however, housing planners had a suitable, comprehensive toolkit of techniques for characterising, investigating and understanding the local housing market, would that then solve the problem? Unfortunately, even with more tools available, there is a strong likelihood they would remain unused.13 A second key inhibitor is behavioural rather than technical. Housing plans are likely to remain weak in terms of both analytical rigour and economic content largely because of the incentive structures facing organisations and individuals involved in housing planning.14

230

Housing Economics and Public Policy

Economists concerned with the theory of economic planning have long been clear that a central aspect of any planning problem is ensuring it is in the interests of agents in the economy to behave as the centre requires them to behave (Heal 1973). Equally, public sector economic theorists have long recognised the issue of appropriate incentives (Niskanen 1994), as in a practical sense do successful practitioners of strategic management (Simpson 1998). Strategic planning of local housing markets clearly has a number of more or less self-evident beneﬁts, but these seldom accrue directly to those individuals and organisations to which the costs of such planning falls. Goodlad (1993) maintains that the beneﬁts of good housing planning include:

•

maximisation of the effectiveness of public expenditure on a long-lasting good, with an ability to address problem causes rather than symptoms;

•

contributions to ‘the democratic process’ as explicit statements of priorities and objectives can be disagreed with and challenged;

•

a feedback mechanism whereby outcomes can be evaluated and future policies made more effective; and

•

a clear guide to action, ensuring resources are devoted to programmes in line with priorities.

Goodlad concludes that, given these advantages, it is surprising to encounter resistance and difﬁcult to understand failure to engage in comprehensive rational planning. But is it really? Firstly, the amount of resource a local authority is willing to devote to comprehensive planning will be conditioned by the extent to which that inﬂuences subsequent resource allocation. The early history of housing planning in Britain was blighted by disillusionment caused by lack of a link between plans and resources. In the English context, Bramley et al. (1999) argue an upswell of interest circa 1990 had much to do with a central government shift in resource allocation, from one based predominantly on a ‘top down’ statistical formula (the ‘General Needs Index’) to one more driven by the quality of authorities’ strategy and performance (Bramley et al. 1999). But as we have seen, the quality of plans produced subsequent to this change was still poor. In the Scottish context, while the allocation criteria were never very explicit, it was known throughout the 1990s that the quality of housing plans played some substantial part in ﬁnal allocations of permission to spend on housing. Again, however, the comprehensiveness of plans fell considerably short of the ideal.

Economics and Housing Planning

231

While plans might inﬂuence the allocation of some resources, they might not be perceived by local authorities to inﬂuence resource allocation enough (in terms of either the total size of the cake or its distribution among authorities) to be worth the associated costs. Blackaby (2000) reports that in some cases authorities may have been unwilling to undertake a comprehensive assessment because even with a better understanding of problems there would be few means available to put them right. A second point is that central government has been consistently unwilling to insist that authorities use any available resources speciﬁcally for strategic planning purposes, and any approaches by authorities for additional resources for such purposes have been rejected. This position must have sent confusing signals to authorities regarding the importance of strategic planning for housing. More generally, while there is no study evidence available to corroborate it, anecdotal evidence suggests that most authorities do not have recruitment policies geared to attracting quality research staff or strategic thinkers. Nor do they have promotion, reward and training systems providing a sound underpinning to support progressive improvement in strategic information availability or the comprehensive assessment of housing requirements. The evidence on plan quality certainly supports this conclusion. As Simpson (1998) notes: ‘if you want “A” but reward “B”, chances are pretty small you are going to get “A”. Employees are very smart – they pay attention to what gets rewarded, not to what gets said.’ (Simpson 1998, p. 626) The issues raised above suggest that mindsets towards housing planning in the UK need to be changed. In fact, ‘the real difﬁculty may be that housing planning as set out in the mid 1970s may now lack credibility and have the wrong emphases. In the advice about housing planning (there was) a presumption that the local authority was in control, that needs estimates could be made convincingly …and that ﬁxed plans would hold over a number of years … The ‘plan’ emphasised a set of ﬁgures, clear investor roles and read as if change and uncertainty were minor considerations.’ (Maclennan 1991, p. 191)

232

Housing Economics and Public Policy

Thus, housing planning as a process has been under-emphasised relative to housing planning as production of a set of numbers and a document. In addition, it has been wrongly presented as something that is comparatively straightforward to do.

Conclusions The rhetoric and the reality of housing planning in the UK differ, and have done so for a very long period of time. Some may ask whether local authorities should continue to have the role that has been assigned them in such an unquestioning way. The equity/efﬁciency case for some form of housing strategy development at the local scale remains a strong one (and not just in the UK). In the absence of any other obvious agency to perform this role, municipal government would appear to remain the best candidate for its delivery, but a new and robust consensus is required about the precise nature of the role, the competencies needed, and the tools and incentive systems both necessary and available to support its delivery. For the future, the development of sound housing planning processes needs more emphasis, and plans produced at a point in time must be understood as contingent on unknown and currently unknowable future events. Modesty is in order. This is not to embrace defeat, nor to suggest that the apathy towards planning for local housing markets evident in the US is a superior model. As Maclennan so nicely (and characteristically) put it, ‘there is no reason why we should restrict our planning to an optimistic disembowelling of sheep in the top ﬂoors of housing departments. For if there are clear objectives, an assessment of the best opening strategic moves and a continuing capacity to observe and react to change, then organisations, with bounded rationality, may best cope with the unfolding future.’ (Maclennan 1991, pp. 191–2)

Acknowledgements I am grateful for the comments provided on an earlier draft by David Donnison, Glen Bramley, George Galster, Alan Holmans, Alison More, Christine Whitehead, and Gillian Young.

Economics and Housing Planning

233

Notes (1)

(2)

(3) (4) (5) (6)

(7)

(8)

(9)

(10)

The views expressed in this chapter are those of the author writing in a personal capacity, and should not be taken to represent the views of Communities Scotland. Within the UK, central government has exercised control of local authority functions through a number of territorially-based departments. For the most part this chapter deals with the situation in Scotland and England. In England, the Department of Transport, Local Government and the Regions oversees the activities of English-based authorities. In Scotland this responsibility falls to the Scottish Development Department. The different parts of the country have broadly kept in step over time regarding what has been expected of local government in housing planning terms. I am grateful to Alan Holmans for advice on this section, as it relates to England. Quoted from private correspondence with Alan Holmans. A national ‘quasi-autonomous non-governmental organisation’ (or ‘quango’) set up in 1989. Varady is an American academic who produced his study on a sabbatical visit, with a primary objective of drawing useful lessons for American housing planning systems. The work by Blackaby complements an earlier study by Evans et al. (1999) looking at the private sector coverage of Welsh housing strategies and operational plans. See DoE (1999a) for the details of the latest English national projections available at time of writing; for discussion of different aspects of these numbers and their sub-national breakdown, see Breheny (1999), Jones & Watkins (1999); for the role of economic factors in household formation, see Bramley & Lancaster (1998) and Bramley et al. (1997). Business Strategies Limited (1998) offer an interesting economic-based alternative approach to household numbers forecasting, as does the work of Peterson et al. (1998), which has attracted some ﬁerce criticism from Wilcox (1998). An alternative approach that projects owner occupancy rates for England to 2016 is provided by Meen (1998). However, in doing so, Meen eschews modelling the propensity for households to form from a given population on the basis of economic considerations, essentially because of data limitations. In an interesting series of papers (Jones et al. 2000; 2001a; 2001b), analysis of price and owner-occupier movement dynamics is explored as the basis for understanding the underlying structure of the Glasgow housing market. This builds on work done much earlier but not subsequently capitalised on by authors such as Hancock & Maclennan (1989) and Maclennan et al. (1987). Barlow et al. (1994) in a survey of all local planning authorities in England and Wales found the working relationship between planning and housing departments to be variable. In many cases there was very little liaison. A later study by Bramley et al. (1999) on the basis of a postal survey estimated that between 1991 and 1998 some 350 housing needs surveys had been carried

234

(11)

(12)

(13)

(14)

Housing Economics and Public Policy out in England by local authorities. Half of these had been developed as a joint exercise between planning and housing departments. This latter study was an assessment of the state of the art in local housing needs assessment as applied to estimating social/affordable housing need at the local authority level (rather than explicitly focusing on overall models of the housing system generally). Bramley et al. give a cogent description of how the housing and planning departments might come together for this purpose. The study ﬁndings do not however suggest it is a good representation of what actually happens. While this review found much to applaud in the work being done by individual local authorities, it still concluded that in conducting housing analysis it is typically the case that: • estimates of social renting at a point in time, or projected forward, are not typically integrated with supply data, either in total or disaggregated by property size, or geographic areas; • employment and income data tends to be used as descriptive ‘decoration’ rather than constituting a part of the overall analysis; and • analysis tends to be ﬁxed at the local authority level, ignoring the issues of broader functional markets or internal sub-market structures (Bramley et al. 1999). For example, movement of young households from rural areas to urban, and the reverse movement of middle-aged households, is equally compatible with a steady state symbiotic operation of urban/rural housing markets, closure of dominant industry employment in a rural area and increased commuting as local prices adjust to reﬂect this, or urban growth increasing the pressure on surrounding settlements. The only study I have come across which shows even a dim awareness of these types of issues is that by Nevin et al. (2001). Rothenberg et al. (1991) using US data have developed the notion of ‘quality sub-markets’ as an alternative to using a spatial approach, which addresses some of the problems involved in using the hedonic method, but its applicability has not been explored in the UK context. A friend and colleague has argued the point forcibly as follows: ‘There is a more basic failure to use economics as a framework for thinking about housing planning issues (even down to just recognising the basics of supply and demand), a failure to think of housing as a system and a general absence of strategic thinking…Housing planning could improve a lot if planners had a more sophisticated conceptual framework, never mind operational models’ (A. More, private correspondence). The key question, of course, is what prevents this from happening? One may of course fairly apply this reasoning to the prior problem. That is, the incentive structure facing economists may well have militated against their applying themselves to the production of useful tools for supporting housing planning.

13 The Right to Buy in Britain Nick Williams

Introduction The most contentious and arguably the most signiﬁcant housing policy in Britain over the last twenty years has been the right to buy public sector housing for sitting tenants (the ‘Right to Buy’). Successive Acts of Parliament have given tenants a statutory Right to Buy the dwelling they are renting from a public sector landlord at considerable discounts off the market price. Since the statutory Right to Buy was introduced by the ﬁrst Thatcher (Conservative) government in 1980 over two million tenants have bought their homes, and this has been a major element in increasing owner occupation in Britain from 55% to 68% of private households over the same period. Local authority-owned housing has fallen from 25% of the total stock to less than 17%, although large-scale transfer to other landlords has also made a contribution to this shrinkage. The introduction of the statutory Right to Buy in 1980 was ideologically driven as part of a wider objective by Margaret Thatcher to reduce collectivist provision across a broad range of goods and services. In the late 1970s, public sector housing was seen as a high priority for this process by the Conservative Party, given the large numbers of council tenants who traditionally supported the Labour Party, and the electoral attractiveness of a policy which (it was perceived) could be used to induce council tenants to switch their votes. A statutory Right to Buy was therefore viewed in prospect not only as a potentially effective mechanism for extending market forces on a large scale but also a political weapon to obtain the power to do so (although see Williams et al. (1987) for evidence that the electoral gain for the Conservatives was limited). Legislation followed swiftly after the Conservatives’ victory in the 1979 General Election.

236

Housing Economics and Public Policy

The reaction from academics and housing professionals to the introduction of a statutory Right to Buy was largely hostile and was itself at least partly ideological, founded upon a deeply-held attachment to council housing and unveriﬁed assumptions about the long-term effects of the legislation. It could be fairly said that both the policy itself and the initial resistance to it were not founded upon either robust evidence or a rational and comprehensive assessment of the likely consequences. More recently, the relationship between policy, evidence and research has become more coherent with the introduction of a modernised Right to Buy in Scotland that attempts to rationalise housing subsidies across tenures. No attempt is made here to provide a history of the Right to Buy. Comprehensive accounts can be found elsewhere (see Forrest & Murie 1988; Jones & Murie 1999; Williams 1993; Wilson 1999). The focus of attention in this chapter is the relationship between policy and research, especially the extent to which the latter has inﬂuenced the former. It will be argued that most of the academic research on the Right to Buy has been reactive, and that until the recent legislation in Scotland most of the policy changes that have occurred have been driven by factors other than a rational analysis of the likely wider effects of those changes. The following sections examine the policy objectives behind successive Acts of Parliament and show how a drive to expand owner occupation (as well as to reduce public expenditure) lay behind them. The details of the relevant Acts are described and the research ﬁndings on the consequences of the Acts are summarised. In particular, I show that this research was used by Conservative governments simply to increase the attractiveness of the Right to Buy as sales ﬂagged after an initial rush of purchases. Lastly, an assessment is made of the contribution that academic research has made to the development of policy in this area, including consideration of the recent legislation in Scotland to modernise the Right to Buy.

Policy objectives The most important objective of the Right to Buy for the ﬁrst Thatcher government was, without question, to increase the number of owner-occupiers in Britain. Theoretical arguments could have been used to justify the policy in terms of removing imperfections in housing markets and increasing labour mobility (see for example Hughes & McCormick 1981; and Minford et al. 1988), but the overwhelming weight of the government’s argument for the Right to Buy was ideological. In moving the second reading of the 1980 legislation Michael Heseltine stated the objectives as follows:

The Right to Buy in Britain

237

‘…ﬁrst, to give people what they want, and, secondly, to reverse the trend of ever-increasing dominance of the State over the life of the individual. There is in this country a deeply ingrained desire for homeownership. The Government believe that this spirit should be fostered’. (Ofﬁcial Report 15 January 1980, cols 1444–5) In terms of this single objective, the policy was a huge success. The impact of the 1980 legislation was immediate. Sales were rapid, rising to a peak of over 220 000 in 1982. Sales then halved by 1985 as pent-up demand was satisﬁed, and (as will be seen) the government introduced further legislation to boost purchases. The capital receipts from sales were used to reduce public expenditure and Conservative governments forced local authorities to use these receipts to pay off debt.1 Critics of the policy and indeed neutral observers pointed out however that there could be wider and undesirable consequences of large-scale sales. These were deﬁned in the House of Commons Environment Committee’s Second Report for Session 1980–81, which was devoted to council house sales (Environment Committee 1981). The possible consequences included the impact of sales on meeting need for rented housing through loss of quality rented stock in speciﬁc localities, impacts on the mainstream housing market through an increase in supply, effects on public ﬁnances and the wider economy, and distributional effects between individuals. The Committee criticised the government for being unable to provide robust information on these issues and recommended a programme of research and monitoring to obtain it. Monitoring was to record information on purchasers and dwellings bought, and research was to cover the effects of sales on relets, management and maintenance costs of the remaining stock, and the distribution of income and wealth.

The evidence base for the 1980 legislation The 1980 Housing Act (and in Scotland the 1980 Tenants’ Rights Act) were not passed in a complete absence of data on the likely pattern of sales or possible consequences of those sales; discretionary sales to sitting tenants had been possible for many years and had occurred on a signiﬁcant scale in some local authorities. The evidence from the period of discretionary sales indicated that it was mainly the better houses with gardens on the most desirable estates that were bought, and that purchasers were drawn disproportionately from middle-aged tenants in work with skilled occupations (see Bassett 1980; Beazley 1980; Forrest & Murie 1976; Malpass 1980; Murie

238

Housing Economics and Public Policy

1975; National Council of Social Service 1980; Richmond 1980; Truesdale 1980). This evidence was indicative rather than deﬁnitive however, since discretionary sales by some local authorities were selective. Birmingham and Leeds, for instance, did not offer ﬂats for sale under voluntary arrangements. It was thus difﬁcult to extrapolate from the evidence to the likely effects of a statutory and more extensive Right to Buy. The government was inﬂuenced however, by strongly-put arguments that a statutory Right to Buy would be directly harmful in two speciﬁc ways. First, that it would damage housing opportunities for older people by reducing the supply of sheltered housing. Second, that sales in rural areas would be harmful because of high external demand for holiday and retirement homes, and because of the already smaller proportion of stock in such areas in the rented sector. No systematic evidence was available at the time to substantiate these claims but lobby groups put them forcefully, and the legislation was consequently amended in two ways. First, sheltered housing was exempted, and second, in certain circumstances tenants of council dwellings in rural areas could be denied some of the purchase rights granted automatically to other tenants (see Williams & Sewel (1987) for a fuller discussion). The legislation had been changed, but as a result of political pressure skilfully applied, not substantive evidence.

Subsequent research on the statutory Right to Buy Given the political prominence of the Right to Buy and the fears that had been expressed concerning the consequences of its statutory nature, there was an obvious need and an opportunity for research to contribute to the policy process by monitoring the pattern of sales, and in the longer term by investigating the wider consequences. Detailed guidance on what this research should address had already been given in the Environment Committee’s Second Report as noted above, but not all of these recommendations were followed in the subsequent work that has been undertaken. Research has been undertaken both in-house by government (see Foulis 1983; Kerr 1988; Lynn 1991; MacNee 1993; Scottish Homes 1996) and also by academics. The most signiﬁcant academic contributions have been made by Forrest & Murie (1984a, 1984b, 1988, 1990a), with evidence from Scotland by Williams et al. (1984, 1986, 1987, 1988); Williams & Twine (1992, 1993, 1994); Pawson et al. (1997); and Pawson & Watkins (1999). The research that has been done has generated evidence on the following topics:

The Right to Buy in Britain

239

•

the pattern of sales in terms of dwelling type and location;

•

the socio-economic characteristics of purchasers and non-purchasers;

•

implications of sales for the role of public sector housing within the housing system; and

•

effects of resold dwellings on the mainstream private housing market.

Notably absent, however, has been research on the speciﬁc and quantitative effects of sales on relets in particular localities and on consequent ability to meet the need for affordable rented housing, as deﬁned by the Environment Committee. In the early period after 1980 this was for the simple reason that the consequences would take time to emerge, since those purchasing their council dwelling would probably stay in it for some period thereafter; but as time went by the absence of this research became less excusable. Undoubtedly lack of work here can be linked to lack of funding by government, and the fact that it is a difﬁcult area to research involving long time-scales and a complex multivariate situation, in which loss of relets through Right to Buy sales is only one inﬂuence on opportunities to rent. Given the importance of the issue however, it is particularly regrettable that no body of substantive work has been undertaken. Evidence on some aspects of the wider consequences of sales for the housing system and for the housing opportunities of different types of household was, however, emerging during the later 1980s. Research was revealing not only a non-random pattern of sales by dwelling type, but also by tenant characteristics. In fact, the pattern of Right to Buy sales was very similar to the discretionary pre-1980 sales. Sales have been predominantly of houses on the more popular estates, and to middle-aged tenants in work with no dependent children. Sales to non-working households, lone parents and those in nonskilled low paid employment have been conspicuous by their absence. There have also been geographic variations. Sales have been signiﬁcantly higher as a proportion of the public sector stock in rural areas than urban areas (but have not been high enough to trigger the reduction of purchase rights built in to the 1980 legislation to protect rural stock). There have also been higher sales in the more prosperous parts of the country, especially the south of England, and lower sales in the less prosperous industrial north of England and the deprived parts of urban Scotland. These non-random patterns of sales have certainly had consequences for the housing system as a whole, and for households’ opportunities within the system, but these cannot be evaluated in isolation from other changes occur-

240

Housing Economics and Public Policy

ring to housing at the same time. The selective loss of properties and tenants from the public sector noted above would have had limited signiﬁcance if the properties sold were being replaced, and similar households were becoming tenants to those exiting through the Right to Buy. Nevertheless, neither of these things has occurred, leading to a general ‘residualisation’ of the public sector in which it has shrunk in size and housed an increasing proportion of poorer and beneﬁt-dependent tenants. This in turn led some commentators to argue that the role of public sector housing has changed to that of a safety net for only the poorest households as in the United States. This has not been a generally accepted view (see Clapham & Maclennan 1983; Forrest & Murie 1990b; Malpass 1983; and Williams et al. 1986), but we can say that the housing opportunities of those households who need or prefer to rent have been reduced since 1980 and that the Right to Buy has made a contribution to this process. To reiterate the point made earlier, however, there has unfortunately been no published research on the actual extent to which the Right to Buy has reduced opportunities to rent in speciﬁc locations, especially in rural areas where fears on this have been most vociferously expressed since the ﬁrst legislation in 1980. Sales to sitting tenants not only affect the rented sector; they also increase the supply of housing for sale should the Right to Buy purchaser decide to resell. Evidence is now available on the role resold dwellings play in the housing market, and in particular whether they continue to increase access to owner occupation for low-income households, or widen choice for those already able to enter, or within the tenure. This evidence comes from England (Forrest & Murie 1990a) and Scotland (Williams & Twine 1992, 1993, 1994; Pawson et al. 1997; Pawson & Watkins 1999). Resale purchasers are generally younger than tenant purchasers, and about 25% said that resale purchase was their only route into owner occupation. Perhaps more signiﬁcantly, most said that the resold dwellings widened their choice of property to buy mainly because equivalent dwellings in the mainstream market were more expensive. This most commonly meant being able to buy a larger house rather than a smaller ﬂat. As well as being younger, resale purchasers tend to be in non-manual or skilled occupations and hence differ signiﬁcantly from those households seeking rented accommodation, who throughout the 1990s have been increasingly drawn from non-working beneﬁt-dependent households. The evidence also indicates that very few resold council dwellings are being used as second or holiday homes, one of the major fears of those who were concerned about the effects on opportunities to rent in rural areas.

The Right to Buy in Britain

241

The relationship between policy and research Much of the reaction to the evidence that has been assembled over time has been limited to the initial impact of sales rather than to the longer-term strategic effects; those with strong ideological positions, for or against, have been able to select evidence to suit their position. Government used little of the research evidence accumulated in the 1980s and 1990s in a strategic way to reassess the role of the Right to Buy in the wider housing system. The only policy responses that can be related to research are the greater incentives to purchase that were introduced in the 1984 and 1986 Acts. The Housing and Building Control Act 1984 reduced the residency qualiﬁcation for Right to Buy from three to two years, with a minimum discount of 32% and the maximum discount increased from 50 to 60% in order to boost sales. Forrest & Murie (1988) estimated that these changes increased the number of eligible tenants by over 250 000. Accumulating evidence both from government in-house research and academics also conﬁrmed that the overwhelming number of sales had been of houses rather than ﬂats. Up to 1986 over 80% of sales were houses. The government’s response to this was to legislate further to make purchase more attractive to ﬂat dwellers. The Housing and Planning Act 1986 increased the minimum discount for ﬂats to 44%, with the discount increasing thereafter by 2% per annum up to a maximum of 70% after 15 years. Sales rose in response to both Acts and by 1989 had recovered to almost 200 000. Flats also constituted a larger proportion of sales, rising to 25% in 1990, but houses still made up the overwhelming majority. (The boost to sales was ultimately short-lived, however, and sales fell sharply again in 1990 to 135 000, partly as a result of recession in the economy and high interest rates. Sales since 1990 have fallen further and have bottomed-out at between 50 000 and 70 000 per year.)2 Yet research has also revealed signiﬁcant effects on the housing system and information regarding losers and gainers from the Right to Buy, that were ignored by policy makers. Purchasers have been allowed to buy good quality dwellings at large discounts and many have made large capital gains on resale. The losers have been those on low incomes who cannot attain homeownership even through the Right to Buy, and for whom affordable renting is the only avenue to acceptable housing. Their housing choices have been dramatically reduced as the public sector has shrunk in size and quality. Since it is the wealthier tenants who have bought, sales have been highly regressive. In the late 1980s and early 1990s council tenants had rents 35 to

242

Housing Economics and Public Policy

40% less than market values but could receive discounts on purchase of 50% or more (see Hills 1991; Maclennan et al. 1991a). The wealthier tenant who exercised the Right to Buy would thus receive a larger housing subsidy than the poorer tenant who had to remain as a tenant. Moreover, the redistribution of housing choice implicit in the Right to Buy policy has been random and unintended. Sales receipts have not been used to reinvest in housing, and the conclusion must be drawn that during the 1980s and 1990s British housing policy including the Right to Buy failed to cohere across tenures. Subsidies remained unrelated to income and households’ true housing preferences were obscured because of the unequal subsidy treatment of the tenures. Sales to sitting tenants are not inherently good or bad but in the absence of a strategic context can have adverse and unintended consequences. As conﬁgured by 1980s legislation, they represented an additional subsidy to a selected group of homeowners and led to distortion within the housing market. Within a broader and integrated package of policies it is conceivable that the Right to Buy could play a valuable role. The experience of the 1980s and 1990s was unfortunately different.

The modernised or strategic Right to Buy in Scotland Maclennan had raised many of these points and potential concerns in a 1977 discussion paper. It is worth quoting directly from this paper: ‘Raised in isolation from a particular context of national housing policy and national economic constraints, the council house sales issue is a red herring. Before deciding whether or not properties should be sold there must be a clear statement of government policy for the growth of owner occupation and policy towards the level and distribution of public sector housing subsidy.’ (Maclennan 1977a, p. 2) and further: ‘…it is only within a dramatically revised system of housing subsidy, which is sectorally neutral and income or wealth related, that council house sales can play a useful role in housing policy’. (Maclennan 1977a, p. 13) It was precisely these conditions that were absent when the Right to Buy

The Right to Buy in Britain

243

policy was introduced and extended from 1980 onwards. Devolution of power to Scotland and the establishment of the Scottish Parliament in 1999, with housing as a devolved function, provided the opportunity for more rational arrangements for the Right to Buy to be established north of the border. In 2001 legislation was passed by the Scottish Parliament that attempts to place the Right to Buy in a more strategic and rational context as described by Maclennan.3 Moreover, this legislation in contrast to previous Housing Acts was based on a review of all the existing published research relevant to the Right to Buy in Scotland. Additional research and statistical modelling of likely future levels of Right to Buy sales based on different levels of discount was undertaken by civil service-based research professionals. This review and research was published during the consultation process to which the legislation was subject (Maclennan et al. 2000). On the basis of the analysis and reaction to it, the Executive concluded that the Right to Buy should be retained in Scotland but rebalanced so that it would be a strategical rather than an ideological policy instrument. Four issues were central to this rebalancing process. (1) What was the subsidy attached to the existing Right to Buy and was it tenure neutral? Did it give tenants of public sector housing a large subsidy on purchase in the form of the discount on the market price not shared by other households with similar incomes but in a different tenure? (2) Was it logical that some tenants should have the Right to Buy and others should not? This was an issue particularly because Right to Buy status of tenants of housing associations,4 by 1999, varied according to a range of factors. Housing associations with ‘charitable status’5 had since 1980 been exempt from the Right to Buy. Those without charitable status could ﬁnd their housing subject to the Right to Buy, depending on the subsidy regime under which a property had been developed, or the form of tenancy agreement held by the tenant.6 (3) Would extending the Right to Buy to non-charitable housing associations threaten their ﬁnancial viability through loss of stock and therefore net rental income? (4) What were the local consequences of Right to Buy and would it be possible to balance community and individual interests in speciﬁc localities, especially in areas where Right to Buy sales had decreased the availability of affordable rented housing?

244

Housing Economics and Public Policy

The modernised Right to Buy as incorporated in the Housing (Scotland) Act 2001 has the following features. To equalise inter-tenure subsidies at the mean (Maclennan et al. 2000), the discounts on purchase have been reduced; they start after ﬁve years at 20% rising annually by 1% to a maximum of 35%, with an absolute ﬁnancial cap at £15 000. This compares with the prior conditions, of a two-year qualifying period with a maximum discount of 60% after 30 years (70% after 15 years for ﬂats) and no cap, and is thus a signiﬁcant reduction in the attractiveness of the Right to Buy for new tenants (existing tenants have a preserved Right to Buy under the previous conditions). The modernised Right to Buy was extended to tenants of non-charitable housing associations who at the time of the passing of the Act did not have it, but suspended for ten years, unless an association chooses to opt in at an earlier date. This was designed to protect associations ﬁnancially and give them time to adjust to the new arrangements. In terms of balancing individual purchaser interests with those of local communities needing affordable housing to rent, the legislation allowed for Right to Buy to be suspended for up to ﬁve years (with the facility for subsequent renewal) in areas designated as ‘pressured’. Pressured areas can be designated by local authorities, with ministers’ approval, where it can be shown by authorities that there is an unsatisﬁed demand for rented housing and that the continuation of Right to Buy sales would be likely to exacerbate the shortage. Perhaps most signiﬁcantly from the point of view of the relationship between policy and research, Scottish ministers have a duty under the 2001 legislation to publish within four years of the Act coming into force, and at regular intervals thereafter, evidence on the effects of the Right to Buy in Scotland on local housing systems. The modernised Right to Buy represents a signiﬁcant improvement on its previous form. It has not been without its critics, however. Concerns have centred on housing association ﬁnancial viability implications, despite the measures built into the legislation to allow time for adjustment (Scottish Federation of Housing Associations 2000). Worries continue to be voiced about the effects of the new legislation on the supply of affordable rented housing in speciﬁc localities, as well as the demotivating effects it will have on housing association committee members who offer their services on a voluntary basis. There are also concerns about the effectiveness of the safeguards introduced in the form of pressured areas in that it is not yet clear what sort of evidence will be required by the Scottish Executive to approve applications for pressured area status. It was noted above that a large gap in the available research evidence is the impact the Right to Buy has had on the

The Right to Buy in Britain

245

supply of rented accommodation at local level. This evidence will continue to be difﬁcult and expensive to acquire. The jury will be out in Scotland for some time as to whether the modernised Right to Buy will have the effects its creators hope for. Only time and evidence can answer the questions that have been raised. Nevertheless, what is fair to say is that there has been an explicit attempt to use research to guide the development of Right to Buy into a more strategic policy instrument, focused on issues of housing equity and efﬁciency in local housing contexts. Moreover, it is likely the situation in Scotland will be judged to be preferable to that in England and Wales even by critics of the modernised Right to Buy, in that there are no major proposals to change the Right to Buy south of the border.7

Conclusions The history of the Right to Buy in Britain is not a happy one for those who believe that housing policy should be rational, based on evidence and with the objectives of maximising housing choice and quality in as fair a manner as possible. It began life as an ideological instrument to expand homeownership, which could be used as an electoral weapon, and rapidly became the single most effective revenue raiser for a government intent on swingeing cuts in public expenditure and personal taxation. The accumulating evidence on the Right to Buy was cherry-picked by the Thatcher governments and used to expand its effectiveness as a consumer policy and revenue raiser divorced from any concern for or interest in the wider consequences ﬂowing from it. It is not as if there was no evidence that could have been used to place the Right to Buy within a more rational and strategic context, even in 1980. Most of the research done on the Right to Buy has been quality work. Unfortunately, it has also been selective and left at least one important gap. Much of the work instigated by academics related to the pattern of sales by property and household and conﬁrmed that the best properties were more likely to be bought by more prosperous households. The possible adverse consequences of these patterns were often averred in terms of reduced opportunities to rent in the social rented sector, but the follow-up research to test these predictions has never been done. It was not until 1990 that the ﬁrst evidence of the impact of Right to Buy resales on the wider owner occupier market was published by Forrest & Murie, and there is still a lack of hard evidence on the effects of Right to Buy sales on opportunities to rent in the social rented sector although this was ﬂagged up by the Environment Committee as desirable in 1981.

246

Housing Economics and Public Policy

Government can be criticised for not funding this research, but academics can also be criticised for not arguing for it to be done. Not until the reform of the Right to Buy in Scotland was all the research evidence available taken seriously (and indeed further work done on prediction of the effects of sales on the rented stock) and used to reconstruct policy.8 Politicians through their control of funding have a major inﬂuence on what research is done, but academics have a duty to deﬁne gaps in knowledge and attempt to ﬁll them even when the results might cast doubt on their own ideological predilections.

Notes (1)

(2)

(3)

(4) (5) (6)

(7)

Between 1980/81 and 1985/86, housing privatisation (mostly Right to Buy sales) yielded over £9 billion, more than half of all privatisation proceeds of £16.4 billion (Forrest & Murie 1988). This was immensely useful to a government that also had as an overriding objective the reduction of personal taxation and the shrinkage of the public sector. The cash raised from council house sales could have been used in various ways including investment in the housing stock, but in fact was mainly used to reduce the taxation burden (see Murie 1989). It should be noted that the 1986 Act was not primarily concerned with the Right to Buy, but with a more comprehensive approach to the privatisation of public sector housing involving block sales to developers for onward sale and large-scale transfers to housing associations (private not-for-proﬁt landlords in receipt of public subsidy whose major objective is the development and management of quality housing for low income households). The declining level of Right to Buy sales after 1990 is an indication that the government had also made a shrewd assessment in 1986 that, in terms of their primary objective to reduce the size of the public housing sector, additional mechanisms would be required. It is not irrelevant to this that the First Minister in Scotland had employed Maclennan in 1999 as a special adviser with a particular responsibility for the proposed Scottish housing legislation! Not-for-proﬁt landlords – see Note 2. A legal status, which circumscribes the operations of housing associations that seek it, but carries certain tax advantages. It should also be noted that where an association acquired stock through transfer of property from another landlord (usually a local authority) and tenants held the Right to Buy under the transferring landlord, that right was preserved under the receiving landlord. There has however been some recent change in England, in that maximum discounts have been introduced which vary regionally, being highest in London (£38 000) and lowest in the Northeast (£22 000). More change may be in prospect.

The Right to Buy in Britain (8)

247

It is tempting to say that Maclennan’s employment as an adviser to the First Minister was crucial, but it must also be said that the housing minister responsible for the Bill, Wendy Alexander, was committed to a reformed Right to Buy and without her enthusiasm it probably would not have happened.

14 The Political Economy of Housing Research David Donnison & Mark Stephens Introduction In this essay we ask: what has shaped the evolution of housing research in Britain; how do we choose and formulate the questions to be studied? And what can today’s housing researchers learn from this discussion that may help them in the years to come? We deal with work by economists but do not conﬁne ourselves to their discipline. Our more fundamental purpose is to remind our colleagues who think of themselves as analysts of housing problems and policies that we should occasionally step back and recall that we are applied social scientists for whom our own profession is also a proper subject for study. We start by looking brieﬂy at the evolution of housing research since World War II, focusing mainly on the ﬁrst and last decades of that period, and making liberal use of hindsight. After this brief review we pause to note developments that have changed the power structures within which researchers operate in Britain and the academic context for their work. These developments help to explain changes in the scale and character of the research done in this ﬁeld. We conclude by drawing a few lessons from our discussion for housing researchers looking to the future. The quantity of research done by academics on housing and the variety of questions they explore have both increased, but there has also been an increase in the competing enterprises doing such research. The comparative advantages remaining to academics working within this wider industry are their opportunities – and obligations – to gain a broad understanding of the workings of their society, and to speak uncensored truth to power and to the

The Political Economy of Housing Research

249

public at large. Since Duncan Maclennan has been a leading exemplar of that philosophy this book is a ﬁtting setting for our argument. Our chapter, we must stress, is not a literature review. Although many authors are mentioned, they are chosen, not to introduce speciﬁc works, but to represent themes, trends and disciplines that play a part in the story. Readers can readily look up their works if they wish to explore them. We give references only to a few of them, that for various reasons should be noted, or call for special comment.

Housing research resumes after World War II Housing research during the interwar period focused particularly on bad housing in urban slums and rural areas, and the effect this had on the health of those who lived there (see works by R.C.K. Ensor, W.G. Savage, Seebohm Rowntree, Marian Bowley, J.M. Mackintosh, and Margaret Bondﬁeld). That emphasis reﬂected both the human situation and the political assumptions of the time. Housing conditions were indeed so bad that they often threatened people’s health. The government department responsible for housing was the Ministry of Health. After World War II, there was no time and little need for research. The problems were critical but simple: the main obstacles to progress were the urgent need to repair war damage, shortages of building materials and skilled labour, and the precarious state of Britain’s post-war economy. For several years the government felt compelled to restrict output to 200 000 houses a year and to channel most of that building through local authorities. Research took off again during the late 1950s and 1960s, based mainly in university departments of social administration, and funded to a great extent by the Joseph Rowntree Trust (now the Joseph Rowntree Foundation). Most of those who played leading roles in the new wave of research (Lewis Waddilove, Rowntree’s director, and academics such as J.B. Cullingworth, David Donnison, Adela Nevitt, John Greve and Lionel Needleman) were motivated by concerns about the needs of working-class people and the development of a welfare state that would meet those needs. First, housing scarcity and then, slum conditions were the problems to be tackled. Later, attention turned to private rents and homelessness, particularly in London (reported on by the Milner Holland Committee in 1965), and then to urban renewal and the small minorities left homeless.

250

Housing Economics and Public Policy

These researchers hoped that the right to a decent home at affordable cost would become as universally assured as the right to free education and medical care.1 Whether they were members of the Labour Party or not, most of them were spiritually embedded in the intellectual wing of the Labour movement, led by Richard Titmuss, head of the LSE’s Social Administration Department.2 When Labour lost power in 1951 these links with the Party were strengthened. For thirteen years its leaders had to do without civil service advisers and they naturally turned to the universities for help – many of them being former academics themselves (Attlee, Dalton, Gaitskell, Crossman and Crosland were among the more distinguished examples). The British housing research agenda of the time therefore reﬂected the housing problems of the time, viewed through a speciﬁc social, political and cultural perspective. A few researchers (Donnison, Greve and later Roger Duclaud-Williams) did their best to learn what was going on in other countries of Europe, West and East. In Western Europe, under similar political systems and faced with similar problems, the basic priorities of researchers and policy makers were much the same as ours. In countries of the Soviet bloc, where the Communist Party and the state were all-powerful, more attention was being paid to sheer volume of output, and therefore to research on building methods and materials. Engineering and architecture were the dominant disciplines in the big research institutes where most of that research was done. Working on a study for the United Nations in the early 1960s, one of the authors of this chapter was going from one European capital to another to consult the ministries responsible for policies in this ﬁeld. In Moscow the top ofﬁcial in the Ministry of Construction of the Soviet Union’s Russian Republic began his speech of welcome by saying: ‘In the Soviet Union we have solved the housing problem. We have learnt how to stick the big panels together.’ He was not joking. Like most of his colleagues sitting round his boardroom table he was an engineer. From the United States, a country relatively unscarred by war and with an altogether different political tradition, there came a cold blast of economic and philosophical reasoning from forerunners of Milton Friedman and Charles Murray. Richard Muth was prominent among them, but even Robert Nozick, in his famous philosophical defence of anarchy, chose rent control as a prime example of the evils of state intervention (Nozick 1974, pp. 270–271). If British academics tended to treat housing as an arena in which to advance the development of the welfare state, many Americans treated it as an arena in which to attack the pretensions of government: LSE social administration versus Chicago economics.

The Political Economy of Housing Research

251

In Australia, where an American plenitude of land and owner-occupied housing was combined with a European respect for the capacity of the state, there came a creatively different research agenda, placing housing ﬁrmly in an urban and economic framework. (Hugh Stretton was – and still is – the guru of this tradition: see, as examples, Stretton 1971, 1974, 1978; but there were many others such as Max Neutze, Ian Manning, Leone Sandercock, Patricia Apps and Patrick Troy.) Academics have recently made criticism of the state for its crippling divisions between separate, ‘top-down’, bureaucratic ‘silos’, an art form of conference speeches. Their own silos have been just as damaging. The institutions and disciplines that at this time provided a base for housing research also restricted the questions the researchers asked. It was difﬁcult to bring the perspectives of social and urban policy together in Britain owing to accidents of academic history. Here, the study of social policy originated in the training of social workers, and recruited from students of sociology and psychology. Meanwhile the study of urban policy originated in the training of town planners, and recruited from students of geography and economics. These two kinds of department usually operated not merely in different disciplines but in different faculties. It is no accident that some of the most distinguished work that broke out of these disciplinary pigeon-holes took place outside the academy (Michael Young’s team at the Institute of Community Studies, including Peter Willmott and Peter Marris, were outstanding examples). As for building and transport, which in countries of the Soviet bloc were researched along with housing and architecture in the great state institutes, these, in Britain, were mainly studied outside the universities in places like the Building Research Station and the Transport and Road Research Laboratory, which recruited scientists, engineers and mathematicians for the purpose. Neither the academics working in social policy departments nor those in the urban studies ﬁeld had much to do with them. We were, and still are, divided by C.P. Snow’s ‘two cultures’. Meanwhile, at this time, no one provided any training for housing managers. Most of them worked for the lower tier of district authorities that had no tradition of training for staff with such modest status – people who had probably worked their way up after years spent as clerks or rent collectors.

Recent research on housing By the end of the twentieth century the housing research ﬁeld had grown considerably. Tables 14.1 and 14.2 show the results of a bibliographical

252

Table 14.1

Analysis of housing articles published 1990–2001 (numbers). 1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

Total

1 14 70 19

0 13 69 26

0 23 68 28

0 26 44 20

3 62 65 37

3 65 75 33

5 90 76 33

1 50 74 27

0 68 115 40

1 87 88 35

2 89 91 59

2 94 95 48

18 681 930 405

10 15 4 9 3 0 10

11 14 1 8 9 0 9

7 25 7 2 7 0 5

9 27 1 7 3 0 7

30 37 0 14 5 1 9

25 59 5 3 3 3 21

35 81 8 7 15 1 20

25 70 9 6 16 0 14

38 90 4 8 20 0 25

72 83 10 5 20 4 25

77 103 12 5 24 0 30

88 113 9 8 26 3 37

427 717 70 82 151 12 212

6 0 2 6 0 0

6 0 3 2 3 1

6 0 3 5 1 0

6 2 1 4 0 1

15 0 7 13 0 1

18 0 8 13 0 2

19 0 12 5 1 0

8 0 9 7 0 0

12 0 8 5 0 0

25 0 13 7 0 0

37 0 11 14 1 0

40 0 9 24 2 0

198 2 86 105 8 5

4

3

1

1

1

0

1

0

1

0

1

4

17

3 0 1 13 0 190

2 0 1 13 0 194

6 0 1 8 0 203

8 0 0 8 3 178

5 0 2 0 1 308

2 0 3 3 5 349

3 0 2 4 1 419

4 0 3 1 3 327

4 0 0 11 0 449

4 1 3 12 5 500

4 2 4 22 3 591

3 0 3 22 3 633

48 3 23 117 24 4341

Housing Economics and Public Policy

Social rented housing/sector Social housing Homelessness Owner occupation or home ownership Housing + economics/economy Housing + land/planning Housing + gender Housing + old people/elderly Housing + ethnic minorities/race Housing + disability/disabilities Housing + labour market/ mobility Housing + health Housing + fuel poverty Housing + construction industry Housing + design Housing + careers/pathways (Social rented) housing + private ﬁnance Private rented sector/ private landlords Housing + affordability Housing + area regeneration Housing + regeneration Housing + subsidies Housing + public expenditure Total

1990

Table 14.2

Analysis of housing articles published 1990–2001 (percentages). 1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

Total

1 7 37 10

0 7 36 13

0 11 33 14

0 15 25 11

1 20 21 12

1 19 21 9

1 21 18 8

0 15 23 8

0 15 26 9

0 17 18 7

0 15 15 10

0 15 15 8

0 16 21 9

5

6

3

5

10

7

8

8

8

14

13

14

10

8 2 5 2 0 5

7 1 4 5 0 5

12 3 1 3 0 2

15 1 4 2 0 4

12 0 5 2 0 3

17 1 1 1 1 6

19 2 2 4 0 5

21 3 2 5 0 4

20 1 2 4 0 6

17 2 1 4 1 5

17 2 1 4 0 5

18 1 1 4 0 6

17 2 2 3 0 5

3 0 1

3 0 2

3 0 1

3 1 1

5 0 2

5 0 2

5 0 3

2 0 3

3 0 2

5 0 3

6 0 2

6 0 1

5 0 2

3 0 0

1 2 1

2 0 0

2 0 1

4 0 0

4 0 1

1 0 0

2 0 0

1 0 0

1 0 0

2 0 0

4 0 0

2 0 0

2

2

0

1

0

0

0

0

0

0

0

1

0

2 0 1 7 0 100

1 0 1 7 0 100

3 0 0 4 0 100

4 0 0 4 2 100

2 0 1 0 0 100

1 0 1 1 1 100

1 0 0 1 0 100

1 0 1 0 1 100

1 0 0 2 0 100

1 0 1 2 1 100

1 0 1 4 1 100

0 0 0 3 0 100

1 0 1 3 1 100

253

1991

The Political Economy of Housing Research

Social rented housing/sector Social housing Homelessness Owner occupation or home ownership Housing + economics/ economy Housing + land/planning Housing + gender Housing + old people/elderly Housing + ethnic minorities/race Housing + disability/disabilities Housing + labour market/ mobility Housing + health Housing + fuel poverty Housing + construction industry Housing + design Housing + careers/pathways (Social rented) housing + private ﬁnance Private rented sector/ private landlords Housing + affordability Housing + area regeneration Housing + regeneration Housing + subsidies Housing + public expenditure Total

1990

254

Housing Economics and Public Policy

(‘BIDS’) search for articles on 18 housing-related topics published in social science journals between 1990 and 2001.3 The tables do not include every article on housing published during the period, and the same article may be counted more than once. Nevertheless, the results are interesting. The volume of published work exploded in the 1990s, rising nearly every year to climb from fewer than 200 entries in 1990 to more than 600 in 2001.4 At the same time, as we shall see, housing research has addressed at least some of the silo problems evident in the 1960s and 1970s. Does this imply that all is now well in the world of housing research? The long-familiar emphasis on social housing and homelessness has continued. Articles on homelessness account for more than one-ﬁfth of those found in our BIDS search – more than 900 over a dozen years – while social housing accounts for nearly 700. There is much more work than there used to be on various aspects of home ownership and its inﬂuence on the wider society – especially its economy. Much of this work was prompted by the disruptive impact of the housing market on the economy in the UK and Scandinavia following ﬁnancial deregulation during the 1980s and 1990s. John Muellbauer’s work on this theme has been particularly revealing and inﬂuential (see Muellbauer & Murphy 1997). So has that of Geoff Meen on the treatment of house prices in econometric models (Meen 1993). Articles on housing and the economy have found their way beyond the housing journals into mainstream economics journals, with a more recent emphasis on the impacts of the European single currency appearing in both.5 The application of economic principles to the study of housing ﬁnance has been a persistent theme since the 1980s, with the two Duke of Edinburgh enquiries reﬂecting these principles. The distinguished work by John Hills on the interaction of housing subsidies, taxation and social security beneﬁts has now spanned almost two decades. Work co-ordinated by Duncan Maclennan in the early 1990s applied survey techniques to the new world of housing ﬁnance deregulation, equity withdrawal and negative equity (Maclennan et al. 1997). The inﬂuence of housing on savings, investment, consumption, labour mobility and labour costs also has been researched, although not as thoroughly as it should have been, notwithstanding Andrew Oswald’s (1996, 1999) interesting, if controversial, work on unemployment and tenure (recently challenged by Coulson & Fisher 2002). Given the problems British society perceives itself facing, there are other, perhaps bigger gaps in the coverage of recent housing research work. There has been relatively little work on housing and health (although the number of articles devoted to this theme grew considerably in the second half of the

The Political Economy of Housing Research

255

1990s). There have been surprisingly few articles about the elderly and people with disabilities.6 The huge importance for housing policy of predictable demographic growth among the oldest people, the rising proﬁle of disability rights and recent legislation on this subject in the UK and the USA all suggest that more work is needed in these ﬁelds. The private rental sector remains under-researched. There has been very little research on the house building industry, and still less on aesthetics and what people want their housing and the neighbourhood in which it stands to look and ‘feel’ like.7 Moreover, with building design and technology on the brink of major changes, social scientists have a part to play in projects now being launched by environmental pioneers to create structures that leave a minimal footprint on the planet. Housing careers, the ways in which people ﬁrst set up home and then move in the course of their lives through housing of different kinds, the capacity of different kinds of cities to welcome newcomers and give them scope for ‘successful’ careers, though not wholly neglected, are important, under-researched themes. So are the careers of houses – the ways in which they are used, extended or subdivided, re-equipped, repaired or neglected over their lives (not surprisingly perhaps: both themes are difﬁcult to research). While there is therefore a much greater volume of housing-related research being undertaken and published, there is still nothing like complete coverage of the issues that currently matter. Some issues have remained mainstays of the research industry; others have emerged, but many important questions remain consistently unexplored. Some would say that the increase in the volume of research work conducted in the academy has been accompanied by a reduction in the real impact this work makes on society. While these developments reﬂect the changing interests of housing researchers, they are also shaped by political, cultural, technical and other factors, to which we now turn.

A changing environment for research Our analysis of changes in the environment in which research has been done starts with the British political setting for research, and then deals with changes in academic resources and priorities. We focus particularly on the 1970s and 1980s – the decades between the two periods examined in previous sections of this chapter, but more recent developments are also brieﬂy covered.

256

Housing Economics and Public Policy

Political changes In 1979 an ice age descended for those raised in the years of hope: academics who believed they had a part to play in carrying housing and urban policies forward in liberal directions, in collaboration with colleagues who worked for the state. Funding for research in most of the social sciences was reduced. The Centre for Environmental Studies, having rashly grown too dependent on government, had the rug pulled from under it and closed. The Central Housing Advisory Committee, which had a distinguished record of researchbased public inquiries into housing going back to the 1930s, was abolished, along with similar Advisory Committees in education and other ﬁelds. The new government had a very simple housing policy: selling council houses, reducing rent controls, relaxing many town planning restrictions – in short, abandoning the ﬁeld. Chicago, for the time being, had won. This attempt to extricate the state from any obligation to provide housing had been tried before by Conservative governments, ﬁrst in the 1930s, then again in the 1950s. But because many people in an unequal society cannot pay for a home that they and their fellow citizens regard as tolerable, only the state can resolve the basic market failure. So, when British governments try to extricate themselves from doing anything about housing, they are always compelled to take up the burden again – usually after scandals and disasters that help to unseat the party in power. Meanwhile, other things were happening to British governance and civil society. The important parts played in policy development by small, independent, London-based groups such as the Fabian Society and – in an earlier generation – the Charity Organisation Society and Toynbee Hall (a middleclass settlement in London’s East End) had long been noted by students of policy development.8 It was inevitable that the deliberate creation of such groups – ‘think tanks’ as they came to be called in Britain – would follow. The term ‘think tank’ originated in the United States to describe research institutes that offered distinguished scholars opportunities for developing and discussing their ideas. In Britain it was given a different meaning, describing research groups of a more partisan kind that have distinctive policy agendas. Here we shall adopt the British meaning but make that clear by escorting the phrase with inverted commas. Viewed from the outside, a ‘think tank’ looks much like a small institute for research in the applied social sciences. But while there is no clear-cut boundary between institutes and ‘think-tanks’ there are important distinctions between those at the opposite ends of this spectrum. Researchers in an institute strive to publish their work in the best peer-reviewed journals,

The Political Economy of Housing Research

257

and measure their success by the respect they win from colleagues working worldwide in their discipline. Researchers in a ‘think tank’ measure their success by the number of their proposals – probably ﬁrst published in pamphlets and in-house journals – that are taken up by the government, and the numbers in their team who are appointed as policy advisers to prestigious ministers. In the heyday of Margaret Thatcher’s regime it was said that the Institute of Economic Affairs had ties made for its staff showing the numbers of times their proposals had been incorporated in legislation. Although we have been assured that this tale is apocryphal, it nicely illustrates the distinction we are making. ‘Think tanks’ have many advantages – particularly for bright, young, childless people who can do without security, for a change of political fashion may consign promising projects to the dustbin and compel those responsible for them to look for other jobs. Being unconﬁned by academic constraints, they can speculate more freely. But their ideas do not have to be put under the harrow of critical review by their peers; nor do they have the daily encounters with life’s harsher realities that the Charity Organisation Society and Toynbee Hall gained from their social work. Thus, disastrous mistakes can be made. The poll tax, ‘back to basics’ (John Major’s slogan for a while) and the privatising of British Rail all seem to have originated at least partly in ‘think tanks’. Their need to remain constantly accessible to power compels them to base themselves in London, dimming their view of more distant parts of the country, and that too may lead to serious errors. Some have argued that supply-side assumptions about the causes of long-term unemployment, unsuited to Scotland, Northern Ireland, the former coalﬁelds and the North of England, are an example of such errors. When it comes to assessing the inﬂuence that housing researchers exert on events we shall have to take the growth of ‘think tanks’ into account. If exerting a direct inﬂuence on government is a researcher’s main aim, an academic appointment may no longer be the best place to start from. There may at the moment be an oversupply of ‘think tanks’ in London, but if Scotland, Wales and Northern Ireland are to develop policies suited to their needs they cannot rely on research based in SW 1 and WC 2: they need their own ‘think tanks’. The inﬂuence of such bodies may have grown, and that of academics been reduced, by the growing importance of radio and television as arenas for public debate and political education, and their producers’ obligations to demonstrate political ‘balance’. They are tempted to stage cock-ﬁghts between speakers who will provide crisp, contentious sound-bites of a predictable

258

Housing Economics and Public Policy

kind – an art form for which partisan policy advisers are better suited than serious academics. Meanwhile the increasing complexity of the analyses expected of scholars in some disciplines and the pressures on them to address, not the laity, but the small, specialised audiences that read the most prestigious peer-reviewed journals, means that they are less likely than their predecessors to be invited to contribute to broadcast discussions. The economists to be heard on the BBC’s Today programme – most inﬂuential of daily commentaries on the news – usually work for consultancies, stockbrokers and banks, not for universities. Its political science, too, is now more often contributed by journalists and polling agencies than by academics. There are more fundamental questions to be asked about the erosion of Britain’s capacity for staging serious public debates about domestic policy. The Royal Commissions, Committees of Inquiry and Advisory Committees we used to rely on for much of this have mostly been abolished. Academics played their parts in these as members, researchers and witnesses. But governments have come to regard the pressure groups, the social scientists and the spokesmen of the public service professions who used to ﬁll their lists of witnesses as a problem rather than a solution. The growing vigour of Select Committees of the House of Commons has gone some way to ﬁll this gap. But politicians’ reports do not attract the attention that was given to a major inquiry like that of the Milner Holland Committee on London Housing or its counterparts dealing with education, health, transport or the personal social services; neither do they generate the same amount of research. Meanwhile the newspaper and publishing industries have changed in ways that tend to stiﬂe academic contributions to debate. Twenty years ago, many of these appeared in New Society, an excellent weekly magazine of the social sciences, in Pelican Books, or in Penguin Specials, which could be got into print within weeks of delivering a typescript. All these have folded or withered. Shorter contributions to debate used to be welcome in the centre pages of The Times and the Guardian, but these and other broadsheet papers now rely mainly on a familiar cast of feature writers whose views become so predictable to regular readers that many give up reading them. Thus changes in government, governance and society in the 1970s and 1980s brought competing ‘ﬁrms’ into the research ‘industry’ in which academics operate, and obstructed the channels through which they were accustomed to communicate their ﬁndings to a wider public.

The Political Economy of Housing Research

259

Change in the academy We turn next to changes in the academic environment itself. Urban studies emerged in the 1970s as a strong suit in British hands, bringing housing and planning research into a broader exploration of the evolution of human settlements, their economies, their social structure, their governance and politics. That development owed a lot to a talented group of scholars, many of them originally geographers (Brian Robson, David Eversley, Peter Hall, Ray Pahl, Nicholas Deakin, Doreen Massey and later Duncan Maclennan) who had escaped from the disciplinary and political allegiances that provided both the inspiration for, and the shackles constraining, much of their predecessors’ work. These people got deeply involved in political and practical affairs – an experience providing the sharpest compulsions to shake off disciplinary blinkers, for you cannot talk sense about policy from within the conﬁnes of one discipline. American scholars also emerged as a major inﬂuence in the urban studies ﬁeld. The Centre for Environmental Studies and Glasgow’s Department of Social and Economic Research played major parts in forging links with them through conferencing, collaborative research and in the columns of the journal Urban Studies. Manchester, the LSE, Birmingham, the Oxford Polytechnic and other academic centres did their best to ensure that Britain could hold its own in this growing ﬁeld. But the Americans, thanks partly to faster adoption of the computer and to a more specialised and focused research tradition, pulled ahead in work of the sort that required rigorous analysis of big datasets. Meanwhile it became increasingly clear, as the best researchers had always known, that housing studies could never be a selfcontained specialism. The computer arrived, enormously enhancing the capacity of researchers to analyse large amounts of data. That increased the power of mathematical methods and the people and disciplines that could use them. (‘It were all ﬁelds when I were a lad’ commented a well-known planner. ‘Now it’s all Forrester and Meadows’ – Jay Forrester, and Dennis and Donella Meadows.) This also gave a stronger hand to the units – many of them in the larger public authorities – that could recruit these people and pay for the increasingly expensive machines they needed. Meanwhile, the run-down of the atomic energy research programme encouraged able physicists and applied mathematicians to move, ﬁrst into transport research and then into the broader urban studies ﬁeld.

260

Housing Economics and Public Policy

Lots of training courses for housing managers sprang up in universities and polytechnics during the 1980s, together with smaller initiatives designed to help tenant activists who were gaining more responsibilities for managing the new, community-based housing associations (Trafford Hall in Cheshire, set up by Anne Power and her supporters, was the biggest centre of this kind). Like academics of every kind, the teachers responsible for these courses were increasingly expected to do some research, which led to more work of the sort that would contribute to their teaching. (Anne Power, Robina Goodlad, Suzie Scott and others led the way in studies of housing management, community-based ways of working, the renewal of run-down estates, dealings with difﬁcult tenants and similar issues.) While these were beneﬁcial developments, mixed consequences ﬂowed from attempts to change management practices within British higher education that, unlike the American system, relied mainly on the state for funds. As the system grew larger and more expensive it was to be expected that the old informal procedures for assessing the performance of academic institutions – relying essentially on gossip among scholars who knew each other and the departments they presided over – would become increasingly inadequate. Laurie Taylor’s weekly column in the Times Higher Education Supplement, tracing the slide into chaos and corruption of one mythical sociology department, cut wincingly close to the bone in some quarters. It was to be expected that more rigorous and bureaucratic procedures for assessing teaching and research would be introduced. These, when they came, powerfully reinforced incentives to do research and to publish it in learned journals while reducing the priority accorded to other ways of using the knowledge gained from academic work – in public debate and journalism, policy advice and voluntary work, evening class teaching and so on. Another change affecting the work of academic policy analysts evolved more gradually. Until the 1950s, university teachers who fulﬁlled their contracts – to teach, administer and do research – found that after working for three or four years on a particular problem, they knew more about some aspects of it than anyone else. If they communicated their ﬁndings in the right ways, politicians and civil servants had to listen to them.9 By the 1970s all this had changed. The growth of the social sciences and their masters’ degrees in universities all over the country, the recruitment of many of our best graduates to new research units in central government and the larger local authorities, and the computers and databanks acquired by these units, together meant that academics could no longer claim to know more than the civil service about anything much. If we were indeed ahead of the game for a few months, the distribution of research summaries and

The Political Economy of Housing Research

261

the data now available on worldwide web sites mean that among ministers’ advisers there are people who know more than we do about most things.

Recent developments In recent years several further trends have emerged that may tend to reinforce cautious orthodoxy, and to make it harder to do independent and innovative housing research in Britain. We note three of these. First, growing pressure on academics to raise their own salaries from research grants and contracts has compelled them to pay growing attention to their funders’ wishes. While much of the research done in this way is perfectly straightforward, the agenda is set by the funder who may also shape the form and timing of publications. Meanwhile, government agencies funding research are expected to ﬁnd internal ‘customers’ to whose needs the work should be ‘relevant’. The Rothschild Report (Rothschild 1971) on the management of publicly funded research and development work was the originator of the doctrine that ‘customers’ must be found for every project, who should be prepared to ﬁnd from their research budgets the funds required for each ‘R and D task’. ‘Customers’ are chosen, not because they suffer from the problems to be researched, but because they control a budget.10 Meanwhile groups with no money to spend on research are of course excluded from the customer role. Second, in recent years interests that once had conﬂicting views about housing policy have tended to converge within a Blair-style ‘big tent’, reformulating potentially difﬁcult social conﬂicts as technical problems. That convergence among organisations speaking for different groups in the housing world exerts a smoothing inﬂuence upon research agendas and publications that may eventually tend to ‘dumb down’ the whole industry in which we are involved. Where genuine conﬂicts of interest or ideology exist we should be trying to articulate and discuss them frankly, not ignore them. Third, Economic and Social Research Council (ESRC) research programmes may be having unfortunate effects. It is entirely sensible that the ESRC, on which so much of the best research depends, should have priorities that are widely understood. These help researchers in different universities to join forces, to focus their efforts more effectively, and to learn from each other. But when these programmes expand to engross two-thirds of the ESRC’s research funds, leaving only one-third for responsive-mode funding, questions must be asked about the inﬂuences that select and shape priorities, the role of the state in that process, and the power of the senior academics who, in

262

Housing Economics and Public Policy

effect, wield patronage within the Council’s committees. With so much research being done on homelessness, no one can claim that the poorest people are neglected, but more subtle inﬂuences may be at work. We should ask ourselves how often we have felt slightly bewildered by the research questions of eccentric younger colleagues – questions that later proved prescient and productive. Would such people get funded today – by committee members who may not for many years have had opportunities to do really innovative research of their own?

The signiﬁcance of these changes What has the upshot of all this been? While housing research has in the main become better grounded in a broad interdisciplinary framework, more rigorous in some respects, more data-driven, and there is more of it about, academics have given contributions to public debate a lower priority. They should remember that the only comparative advantage now remaining to them is their independence and their capacity – if they hold on to it – to speak truth to power. Ofﬁcials should also speak truth to power and have more opportunities for doing so, but academics still have greater scope for picking the questions they study, for consulting whomever they wish, and – if they negotiate well-judged funding contracts – for publishing their ﬁndings. We return below to the implications of all that.

The housing research industry We work in an industry, closely linked as customer and supplier to other industries – those that send us students, pay us to teach them, and employ them thereafter; that fund our research and give us access to the sources it depends on; that publish our books and journals, and buy them. Housing research is also related to other academic industries – the race relations industry, the gender studies industry, the disability industry, and so on – whose practitioners occasionally contribute to our journals and our teaching. What we can do depends on the perceptions and priorities of people in all these industries and our capacity to turn their attention to new issues – to create and market new products. The options for research on housing are now potentially limitless. The subject was once a fairly distinctive ﬁeld, posing a limited number of major questions. Now, every policy problem can be seen to have a housing component. Hospital waiting lists? – can nurses be recruited in areas where housing costs are too high for them? Support for university students? – should they

The Political Economy of Housing Research

263

be living more cheaply at home, now that we have improved their parents’ housing so much and set up universities in every major town? Ozone layer depletion? – can we arrest its destruction without reducing the spread of suburbs and the growth of commuter trafﬁc? Inﬂation? – can we control it without tighter regulation of the mortgage market? In many parts of Britain the basic shortages of tolerable housing that used to dictate research priorities are a thing of the past – although we have cut back output so far that in some places old-fashioned shortages are re-emerging as a problem. But there is now no limit to the things that housing researchers could usefully study, which is a happy but bewildering position to be in. We offer a few suggestions that may help those grappling with this dilemma.

Lessons for housing researchers While we may be compelled to take on a good deal of pot-boiling work to keep ourselves and our colleagues employed, we should never forget that our independence and integrity are the only comparative advantages left to us. So we should beware of being used by those who sponsor research simply to give the illusion of action, to delay decisions, or to gain support for decisions already made. If we do not from time to time ask questions that the state and other corporate bodies would not pose and produce ﬁndings that they would not publish, then we would do better to work in a consultancy or a government research unit. The mass of work done among homeless people in recent years suggests that no researcher should be prepared to do any more until a major programme of action to help them creates a new situation that may be worth studying. The most important thing that academics can do is to extend general understanding of our society, the way cities and regions work, and the opportunities they offer to different kinds of people. That kind of research, sometimes done by people who have little interest in speciﬁc policy issues, is not likely to be done by the state or the private sector. Policy-focused research of the kinds that government needs has to be embedded in the broader understanding of society that we should be trying to provide. Meanwhile, outside the academy, society needs a properly informed ‘political class’ with access to meeting places where people can debate public policy issues, together with ‘think tanks’, independent of the civil service, where people can shape innovative ideas into a form that governments can use. Academics can play an active part in developing and informing this political class.

264

Housing Economics and Public Policy

Policy discussion must not be conﬁned to the middle-class. We should always remember – what the state too often forgets – that the people who experience a problem are the best single source of evidence about it. We should not only interview them as subjects of our research; we should consult them before going ﬁrm on the questions to be researched, consult them again about our conclusions for policy and practice when they are still at a draft stage, and insist that they be invited to play an active part in conferences at which we present our ﬁndings. It can be done.11 We should bear in mind that the growth of the ‘welfare state’ that spread government inﬂuence into many ﬁelds of housing, health care, employment and other matters hitherto dominated by the private sector and charitable bodies did not bring class conﬂict to an end. It transferred many of the conﬂicts that used to divide landlord and tenant, employer and worker, middleclass and working-class to the counters of the public services where they may be more fairly managed but continue to this day. Every time we launch a research project on social rented housing after consulting its owners while failing to consult those who live in it, every time we set up a new course for housing managers without providing any education for those who are to be ‘managed’, we are taking sides in those conﬂicts. Social exclusion is a process. If we are not among the includers then we shall ﬁnd ourselves among the excluders. Despite the emphasis placed on ‘relevant’ research for which ‘customers’ can be found, we should remember that even in policy-oriented research, the most important output of a study is often the education of the people who do it. They will bring what they have learnt to bear in due course, by publishing learned papers, offering policy prescriptions, teaching their students, doing voluntary work, or contributing in various ways to public debate. Speciﬁed outputs and deadlines for such studies should be required; but the longerterm aim of the work is to contribute to a national intellectual and political resource capable of mounting intelligent discussion in public gatherings and through the media, and raising the general level of public debate. That includes creating effective ‘think tanks’ that feed ideas to governments. When it comes to communicating research ﬁndings, the inﬂuence of researchers usually depends in practice not on the authority and rigour of their conclusions or the elegance of their arguments, but on the ‘stories’ they tell about society’s problems, the moral meanings of these stories, and the timeliness and vividness of the tale. Like proverbs, the messages conveyed by these stories often conﬂict. It is the politicians’ job to decide which story is most relevant to the situations they have to deal with.

The Political Economy of Housing Research

265

Some stories may be demonstrably wrong, but it is not enough to prove that a Charles Murray (1984) has made a lot of errors and leave your readers discussing the same problems in the light of the same moral values. Another and more reliable story, formulating the problems in new ways and giving them different moral meanings, has to be equally vividly presented.12 Stories about our society and the way in which it works can be presented with algebra and in ﬁgures as well as in prose. The Treasury’s model of the British economy is one such story. Note the characters that play a part in it: incomes and taxes, savings and investment, imports and exports … all of them obviously important. But health, happiness, security, solidarity, education, respect for authority … none of these appear in the cast, and they may be more important for the nation’s welfare. There are researchers who are developing new descriptions of social welfare, how it is created and how different countries can be compared, who are bringing these characters on stage (see, for example, Marmot & Wilkinson (eds.) 1999; Doyal & Gough 1991).

Priorities for Scotland Since many of our readers will be in Scotland it is appropriate to illustrate some of the practical conclusions to which these arguments lead by offering some thoughts about this country. In this ﬁeld as in any other, Scottish researchers could doubtless do better, and we have suggested gaps in their agenda that they may want to ﬁll. But they have a more creditable record of research and teaching in the housing and urban studies ﬁelds than their colleagues in most parts of the world. Urgent at this point in Scotland’s history is the development of lively and well-informed political debate that involves people from all classes in all parts of the country, and the creation of two or three Scottish ‘think tanks’ to help to form links between the academy, the state and a wider public. The Institute of Contemporary Scotland is trying to create arenas for betterinformed public discussion. It brings different kinds of people together to debate urgent issues, sometimes at short notice, in different parts of Scotland. Other bodies are doing likewise: the Church of Scotland and its Committees are among them. Meanwhile the Scottish Parliament is endeavouring to work in a more open, participative, accountable fashion than its Westminster counterpart. The Scottish Council Foundation, recently declared to be the ‘“think tank” to watch’ in an appraisal of all those operating in the UK,13 is making an impressive attempt to ﬁll the second role. Housing researchers should be contributing as opportunities arise to discussions in both these arenas.

266

Housing Economics and Public Policy

In conclusion We conclude by drawing together some of the most important of the points touched upon in a wide-ranging discussion. Academic researchers interested in the housing ﬁeld will ﬁnd that it offers them ways into every kind of issue, not the more clearly deﬁned research agendas their predecessors had. However, their monopoly of intellectual inquiry has long gone. They will ﬁnd many other researchers working alongside them in consultancies, voluntary agencies, pressure groups, ‘think tanks’ and the research units of central and local government. Any comparative advantages the academics may retain now arise, not from superior knowledge, intelligence or data but from the scope they have for choosing their own research questions and publishing their own uncensored ﬁndings. They should use that privilege to gain a broad understanding of society and the place of housing within it. Such studies will provide a sounder framework for, and a wiser understanding of, the more sharply focused, policy-oriented research that they and others will also do. Competing research groups based in the state, the voluntary and private sectors have less opportunity to do this broader, scene-setting kind of work. Researchers should always remember that the people who experience a problem are usually the most important source of evidence about it. Such people should be involved, not only as passive informants, but in formulating research questions, discussing ﬁndings and recommendations, and contributing to public discussion of the issue. Academic researchers should take whatever opportunities they may have for contributing to – and learning from – public education and policy discussions. In doing so they will help to build the penumbra of intelligent and well-informed citizens that every department of government needs if it is to work effectively. They will ﬁnd that in the longer run their own inﬂuence in these discussions depends on helping people to see their society in new ways by offering them stories with moral meanings that clarify the character of its problems and suggest better ways of addressing them.

Acknowledgement We thank Tony O’Sullivan for helpful editorial advice. We are also grateful to Emily Lynch for processing the BIDS search presented in the tables.

The Political Economy of Housing Research

267

Notes (1)

(2)

(3) (4)

(5) (6) (7)

(8) (9)

(10)

(11) (12) (13)

The Housing (Homeless Persons) Act of 1977 for a while came close to providing that assurance, at least for married couples, households with children and elderly people. An account of the teaching provided in Departments of Social Administration throughout Britain can be found in Donnison (1961). Most of the research going on in these Departments took place at the LSE. We could have gone further back in time but feared the record would be increasingly incomplete for earlier years. Housing journals now attract more and more work – in humblingly good English – from other countries, particularly the Netherlands and Scandinavia, besides a steady ﬂow from the USA and Australia. Authors from the Far East appear fairly often, but there are fewer writing from, or about, third world countries and the former Soviet bloc, although contributions from both are growing. For example, Maclennan et al. (1998b); Memery (2001). Although again the small number dealing with ethnic minorities grew in the second half of the 1990s. Despite studies made in earlier years by Pearl Jephcott, Kevin Lynch, and Barbara Adams’ team at the Department of the Environment, applied social scientists have generally dismissed aesthetics as ‘unresearchable’, or – worse still – condemned such work as ‘environmental determinism’. The Sunday newspaper supplements and women’s magazines however have never hesitated to write about aesthetics, and, at the end of the day, it was often what the buildings and their surroundings looked and smelled like that most discredited public housing and municipal socialism. Hall et al. (1975). See also Donnison (2000). ‘We pay no attention to the books you people write,’ said a senior housing ofﬁcial to one of the authors in the early 1960s. ‘No-one in government has time to read books. But if you write an article for The Times or the Guardian I have to read that because the Minister will have seen it before he gets to the ofﬁce in the morning, and I have to have the answers for him.’ To be fair to him, Rothschild made it clear that he was not talking about the social sciences – but the government did not hesitate to apply his doctrines in this ﬁeld too. ‘I was only trying to tell them how we did it in Shell,’ he said when his recommendations were criticised. These are sensible doctrines for technological research of the kind that a big oil company needs, where problems and solutions can be clearly speciﬁed. The study reported in Bannister et al. (1993) was handled in this way. Marris (1997) provides a revealing analysis of these processes from which much of our own version is taken. Prospect/Nirex awards; Prospect, November 2001 p. 7.

15 Policy and Academia: an Assessment Richard Best

Introduction To what extent can academics play a part in national policy-making? This chapter considers the extent to which published research from academic institutions, and the input of individuals from the academic community, can inﬂuence government ministers, civil servants and others responsible for formulating social policy. It concludes that the contribution of research can be of considerable importance in determining changes to policy. But the transmission and translation of the ﬁndings from research require special skills either on the part of the academics themselves (for which they are seldom trained or supported) or through the use of skilled intermediaries. The chapter ends with suggestions for enhancing the policy-making role of academia.

Separate worlds? This chapter concentrates on social policy, with particular reference to housing issues, and on the academics, and the research studies undertaken outside government, that seek to address social policy questions. Those who approach policy issues from an academic perspective, and those whose job it is to make policy decisions, sometimes appear to inhabit different worlds. To policy-makers, academics always seem to look back at what has gone before. Researchers usually appear to be in no hurry and, after assembling the evidence, seem loath to come to ﬁrm conclusions about the changes to current policy that would make a difference.

Policy and Academia: an Assessment

269

By contrast, government ministers face constant pressures, scrutiny and criticism. They must keep looking ahead (not least with an eye on the potential popularity of any changes they might announce). They are in a hurry – after all they may not occupy their posts for very long – and they are likely to have clear and ﬁrm opinions. While academics may be comfortable with the conclusion that ‘more research is needed’, policy-makers must take decisions and must take them quickly. When the Joseph Rowntree Foundation presents ﬁndings from research it has commissioned, the response from politicians often goes much like this: ‘You have drawn our attention to shortcomings and problems which have resulted from previous policies. But you have said very little about what would work in the future. We have to take decisions now, in the real world, with all its ﬁnancial and political constraints. And we need clear guidance. We do not want to hear from those who can criticise but cannot do.’

Mutual interests While policy-makers may accuse academics of an inability to make up their minds – ‘the paralysis of analysis’ – academics may be justiﬁed in criticising politicians for acting before all the facts are known – ‘jumping to conclusions’ or ‘shooting from the hip’. But despite the apparent divide, those within both groups who want to change the world for the better must grudgingly admit that they need each other. Even though the primary concern of politicians may appear to be the gaining or retaining of power – which means taking decisions on the basis of their electoral popularity – most recognise that they need a knowledge base on which to take decisions. They need to understand the origins, the causes, the dimensions of whatever social problems come within their remit and they need to get advance warning if their proposed policy changes may lead to failure. Even when politicians appear disdainful of academia, most are hungry for the information that will give them an edge, keep them in the vanguard of new thinking and justify the decisions they want to take. Academics may see politicians as operating on a superﬁcial plane, pandering to the electorate rather than going where the evidence points. But it takes a different temperament to move beyond the comfort zone of clear evidence. Politics involves risk-taking and a career in the public eye that can be swiftly terminated by the democratic process or by the judgement of superiors. The academic life seldom brings fame or fortune (and it can be as insecure as politics for those not yet in its senior echelons), but it has other attractions.

270

Housing Economics and Public Policy

While politicians are constantly required to face in several directions and must often avoid the slings and arrows not just of opposing parties but of their ambitious colleagues, those who make a career in academia can ﬁnd it easier to retain their integrity.

From searching to changing The Joseph Rowntree Foundation commissions and supports a very large amount of academic research activity.1 It does so not because it wishes to sustain the academic community but because it believes its underlying mission is most likely to be fulﬁlled by putting high quality research into the hands of key policy-makers and practitioners. At the beginning of the twentieth century, Joseph Rowntree set aside substantial resources to ‘search out the underlying causes’ of social problems. He believed that if only the truth could be discovered about the root causes of poverty and disadvantage, the right policies could be put in place to overcome these social evils. He had seen the ways in which disease and death had resulted from outbreaks of cholera and typhoid; and he had seen how public health legislation had treated the problems at source and conquered them. Could not the power of science lead to policy changes that would vanquish social ills just as they could overcome physical ills? At the Foundation that Rowntree created, the emphasis today remains on trying to secure policy change through funding academics to discover those ‘underlying causes’. But, standing between the government machine and the research community, we have had to face much frustration: sometimes academia cannot shed light on the issue because the evidence is absent or the policy question cannot be answered by a research project. Sometimes research does indeed provide grounds for believing – if not on the basis of certainty, at least on probability – that existing or proposed political action will be positively harmful, and yet politicians brush the evidence aside. Sometimes the position is reversed and it is the inability of researchers to operate within tight timescales – or to respond positively to the politician’s need to move from knowledge to action – that means the chance to effect change is lost. However, despite the frustrations that can arise from the lack of connectivity between the output of academia and the demands of policy-makers, I remain convinced of the key role for research in improving social policy. Unless policy is grounded in knowledge and understanding it runs the risk of doing more harm than good.

Policy and Academia: an Assessment

271

Improving the connections between research outputs and policy-making may in general require the presence of some additional ingredients to close the gap between what the researcher produces and what the policy-maker needs. But before exploring this further, this chapter considers how academics themselves may sometimes directly engage in policy formulation.

Crossing the divide Occasionally there are politicians who have been trained as researchers and understand the disciplines involved. There are also academics that walk freely within the policy-maker’s world, with a clear understanding of what it takes to convert research into a policy-making tool. As an academic with much experience of both sides of the line, most recently as Special Adviser to Scotland’s First Minister, Professor Duncan Maclennan is an example of someone who has bridged the gap. His work in helping politicians who need information and insights has demonstrated to other academics that the policy-makers value what they possess. What are the pros and cons of those with research skills leaving their ivory towers for a career outside academia where policy-relevant insights can directly inﬂuence decision-makers?

Academics in Parliament As a relatively frequent visitor to the Eastern European countries, I have noted how often leading academics have gone the whole hog and become government ministers. This can have the disadvantage of the individuals ‘going native’ and forgetting the academic rigour which they should be bringing to their position of power: they can become so caught up in the cutand-thrust of the political system that their ability to exercise any academic discipline can desert them. Academics in the UK enter national politics less frequently, but it does occur. A good example is the Liberal Democrat MP Steve Webb, who was also Professor of Social Policy at the University of Bath.2 There are severe difﬁculties in the UK – even at the local level – of becoming a politician during or after a successful academic career (even if the back-stabbing and naked ambition of party politics is acceptable!). Nevertheless a Parliamentarian with real academic expertise can contribute to the processes of scrutiny, for example through Select Committees where political point scoring is not so endemic, as well as through their input to the legislative processes.

272

Housing Economics and Public Policy

Moreover by retaining links to the world of academia they can act as a bridge in translating and transmitting the politicians’ needs to their colleagues in the research community. Recent reforms to the House of Lords have led to the appointment of a number of academics as independent (‘cross-bench’) members, including the Vice Chancellor of Edinburgh University, Professor Stewart Sutherland, and the Director of the Royal Institution of Great Britain, Professor Susan Greenﬁeld, as well as Professor Sir Robert May of Oxford, and Sir Claus Moser. There are proposals for more appointments to the Lords by an Independent Commission set up for the purpose – so the United Kingdom Parliament could in future see more academics with their hands rather nearer to the levers of power.

Think tanks A half-way house for some of those with research knowledge and an interest in policy lies in engagement with the think tanks that stimulate government thinking from the outside, or in working as an adviser to government on the inside. There is a good deal of crossover between these two: people from think tanks often move into government advisory posts. For example, Geoff Mulgan, the founder Director of DEMOS, moved to the No. 10 Policy Unit after Labour’s 1997 election victory and now heads the Cabinet Ofﬁce’s Policy and Innovation Unit. (Rather more commonly, the think tank director goes on to become a Member of Parliament (or, no doubt, a Member of the Scottish Parliament). Thus David Willetts moved from the Centre for Policy Studies (CPS) to become an MP and then a prominent member of the Cabinet.) The think tanks – such as the Fabian Society, the John Smith Institute, the Scottish Council Foundation, the Institute of Economic Affairs, and the Adam Smith Institute – can be the custodians of a particular ideology, acting as a ginger group for a mainstream party. They can feed that party – whether in opposition or in power – with new ideas (or re-packaged old ones). Or they can be stalking horses for a political party to test out a bold plan and gauge public and media reactions before the party decides whether to absorb it into the mainstream or reject it. When their patron is in power – as with the CPS in the days of Mrs Thatcher’s premiership or as with the Institute for Public and Policy Research (IPPR) today – the think tank will be fêted by those who want to inﬂuence government, or those who want to know where government policy may be heading.

Policy and Academia: an Assessment

273

The problem for academics in becoming too entangled with the think tanks of their choosing is that each has an agenda. Impartiality and independence of thought may not sit well with the vision of the think tank: evidence is needed to support their case, to justify their philosophy, not to expose weaknesses or query their premises. Moreover, to capture the attention of the news media, they are sometimes tempted to be deliberately provocative, expressing opinions that could never be justiﬁed by evidence or objective analysis. No academic can be entirely free from their own prejudices and, indeed, value judgements; but for many the approach of the think tank will run counter to their professional training and their instincts. It may better suit the temperament of the committed campaigner than the dispassionate researcher.

Government advisers The academic who becomes an ofﬁcial adviser moves to the heart of government where sometimes – but not always – there is a hunger for knowledge on which to base new policy. There are risks for any adviser of being seen thereafter as partisan and politically aligned. But in reality the value to those they advise is not (as it is with the think tanks) in public pronouncements, but in private guidance: for this, the important qualities are the retention of an independent mind, a willingness to present research evidence fearlessly, the exercise of critical faculties and the ability to see more than one side to an argument. Career civil servants can feel disempowered by the presence of such advisers. But they can be disarmed by charm and tact, and by the adviser working within the conventions and sensitivities of the system; in-house advisers from academia can then be immensely important – and without the necessity of compromising their intellectual integrity. To fulﬁl this role effectively, some direct involvement in non-academic institutions concerned with policy implementation can be an undoubted advantage. Duncan Maclennan’s career demonstrates this clearly. Prior to appointment as a Special Adviser to Scotland’s First Minister in 1999 he served on the Board of Scottish Homes. Combining an academic career with public service in the housing world has helped many to balance the evidence and insights gained from knowledge of research ﬁndings, against the political practicalities required to get things done.

274

Housing Economics and Public Policy

By serving in capacities outside the university, academics not only contribute directly to policy-making (or implementation) in the real world but they also lose the naivety that those conﬁned to ivory towers can display. Those on the boards of housing associations have found out, for example, how powerful can be the inﬂuence of central government over organisations dependent on a ﬂow of grants and subsidies. Those who serve on local authorities soon discover that the ideal solution may simply be unobtainable because of its cost or electoral fallout. Experience on a health trust or other quango can teach the ability to compromise and concede some ground – not least to accommodate the ambitions and idiosyncrasies of the other personalities involved – and yet to avoid making concessions that really count. Duncan Maclennan exempliﬁes many of the qualities that lead to academics being drawn into the heart of government policy. He is secure in his research knowledge; he is highly skilled in synthesising and bringing together the key points from a body of available research; he can present his material to busy political masters in an engaging and accessible style. But one essential step in gaining the conﬁdence of ministers probably lies in gaining experience within statutory or voluntary bodies where different skills can be honed.

Shakers, not movers What of those who cannot or do not want to move on from their academic base, but who still wish to play a direct, active role in translating research insights into policy or practice change? For some, opportunities may arise from the invitations to contribute periodically to policy-making task forces, working parties and committees. For example, in England, the Cabinet Ofﬁce’s Social Exclusion Unit has called on individual academics – many of whom have been supported in their research by the Joseph Rowntree Foundation – to sit on policy action teams and to help with particular pieces of ongoing work. Another example is the House of Commons Select Committee: many of these have drawn on academics as their advisers in their role of scrutinising the work of government departments. Professor Christine Whitehead, for example, has been a key adviser to the Environment Committee. In collaboration with the clerks to these committees, external academics can prove highly inﬂuential. In Scotland, similar input to the new Scrutiny Committees, even if on a relatively time-limited basis while a particular topic is under investigation, can provide the chance to make very productive use of

Policy and Academia: an Assessment

275

knowledge gained from previous research. Reputation is the key to invitations to contribute to policy in these one-off advisory capacities.

Making the connections Most of the input from academia to the world of policy-making will, of course, come not from direct action by academics, but from policy-makers assimilating the actual work done within the academic base. What are the key factors that will determine the extent of this works inﬂuence on policy? Some of the main ingredients must be:

•

who pays for the research;

•

how it is presented, publicised and disseminated;

•

the quality, relevance and timeliness of its content; and

•

partnership and persistence in taking forward its messages.

Pipers and paymasters Who pays for research will directly affect its potential for inﬂuencing policy. Most social policy research in universities involves an external ﬁnancial input, particularly where it is applied rather than theoretical. But external funders are likely to inﬂuence what is produced (even where they say – and perhaps believe – that their involvement is entirely dispassionate). Whether research is delivered under contract and must accord with the precise requirements of those who commission it, or whether it is initiated by researchers seeking funds from a research council or a foundation, the source of funding will exert an inﬂuence. Undertaking research for government departments can bring the satisfaction of reaching directly into the policy machine. Where government wants to ﬁnd out whether, for example, some pilot scheme has proved a success, a great deal depends on what the research unearths. Moreover, the relationship between researchers and senior civil servants, through the execution of completed work, can lead to more direct opportunities to input to the policy process. Thus, academics with a proven track record in concluding work for government departments may get drawn into those governmental task forces, working parties or ministerial sounding boards. When conﬁdence and

276

Housing Economics and Public Policy

trust is built up through the contractual process, it can lead to invitations to participate more directly in policy processes. The downside from ‘accepting the king’s shilling’ and undertaking research that is paid for by the public purse is that the political priorities of the paymasters can bring tensions and frustrations for conscientious academics. Ministers are unlikely to sanction research which may prove their policies were ill-conceived; when seeking a justiﬁcation for a course of action to which they are already committed, they see research as the route to proving their point or vindicating their actions; and all too often the evaluations of projects they have set in hand will either be ignored or kept out of the public eye if they produce the ‘wrong’ result. Although the funds available from charitable foundations can be stretched much further if their support for research projects goes alongside matching contributions from government departments, at the Joseph Rowntree Foundation we have generally eschewed joint funding: the constraints and compromises inherent in sponsoring work alongside a government partner – which must take account of political expediencies – has proved too restricting for us. For very different reasons, the Foundation has found it difﬁcult to co-fund research with the Economic and Social Research Council (ESRC). The ESRC’s aims are to ‘identify, support and sponsor the highest quality of social science that Britain can produce’; this may or may not connect to policy change. The very factors that give much-needed security to academics – guaranteed support over a relatively long period, without severe pressures to produce speciﬁed outputs within tightly deﬁned schedules, and the opportunity for a more reﬂective, less narrowly focused, approach – may produce results that lack the immediacy and relevance that the policy-makers demand. The purpose of ESRC funding may be more directed to ensuring a corpus of research capacity, and backdrop of knowledge and data, than to achieving policy-relevant outputs of immediate and direct use to decision-takers. And what of the role of foundations, charitable trusts, that support academic research? Confusingly, no two are alike. Each has its own special interests, perhaps relating to the wishes of its founder or the interests of its current trustees. Some have neither the inclination nor the resources to do more than send cheques to worthy academics to get on with the work they have proposed; others interest themselves deeply – with staff to pursue their engagement – in every stage of each research project. The attribute they are likely to have in common is that they stand outside the imperatives of politics, where popularity must be courted, and of business, where research

Policy and Academia: an Assessment

277

must contribute to the organisation’s proﬁtability. Their independence is their common feature but sometimes it seems that the decisions that ﬂow from it are whimsical or simply mysterious. The Joseph Rowntree Foundation is typical in some ways and distinct in others. It, perhaps more than most, reacts against the notion of simple grant-making: we have rejected ‘cheque-book charity’, where the objective is accomplished when the grant is made, in favour of a partnership with the academic body concerned. A partnership in which one side supplies all the money and another does the majority of the work can seem lop-sided. The onus on every foundation is to avoid an abuse of this potentially powerful position: we have to ensure that, even in subtle ways, we are not choosing and guiding research that will conﬁrm any prejudices and unproven opinions which individually or collectively we hold. The value of research we support is undermined if it is not always searching for truth. While we may be helped by the integrity of the Quaker tradition we inherit, we must constantly guard against suggesting, let alone coercing, researchers to follow a particular path that appeals to our own sensibilities. The essential fact remains – who funds the research will affect how it subsequently inﬂuences policy-making and policy-makers.

Presentation and dissemination Turning to the ways in which academics can themselves help determine whether their work affects policy, a key component is its presentation. If the intention is to inﬂuence policy-makers – national or local politicians, civil servants, and the decision-takers in quangos, voluntary bodies and the business sector – and to reach those who themselves exert inﬂuence on policy – the think tanks and the news-media – the academic community must avoid simply talking to itself. Rewards – of money as well as prestige – often ﬂow from impressing fellow academics and from contributing to academic journals. Incentives are lacking to present material in the formats that appeal to those with policy responsibilities. This seems perverse if academia – funded by the wider public for the public good – is to be properly encouraged to be as valuable to the processes of public policy as it could be. There must be a danger that, if outputs from research appear to be intended for other academics and are not accessible to those who genuinely want to understand more and learn from the work done, research based in universities will come

278

Housing Economics and Public Policy

to be regarded as irrelevant and self-indulgent; ultimately, this must affect the perceptions of funders and will threaten the support available from governmental and independent sources. If the subject covered by research is intended to reach the desks of policymakers, it must use everyday language, be free from jargon, equations and initials (even if referenced at the back), and from elaborate references to its methodology. Technical appendices can contain the items that interrupt the ﬂow of the ‘story’. Academics must remember that their readers are more used to gaining information from newspapers and magazines and the only lengthy texts they read are likely to be the works of world-renowned authors. Academic texts are an essential part of the process: short reports and summaries only have a raison d’être if there is a full report behind them. But if the only source of reference is a full-length academic book, or an article in a journal unknown to most decision-takers, the chances of policy change will be slim. Unfortunately few funders want to pay for editing, literary criticism and investment in the appearance of the work undertaken by researchers. There may be very little cash available for publicising and disseminating the contents of research reports, e.g. through a public launch, articles in newspapers and magazines, specialist seminars, direct presentations for government departments, and the rest, yet with ﬁerce competition for the attention of busy people who have already far more to absorb – now on the internet as well as through the pile of documents that come in the post – important research is likely to go unnoticed without investment in its presentation and dissemination.3 An illustration of such investment and the beneﬁts that ﬂow from it comes from the dissemination work that supported a programme of research conducted at the beginning of the 1990s on aspects of housing ﬁnance. A series of reports – several based on a survey of 10 000 households in six conurbations with accompanying case studies – were produced for the Joseph Rowntree Foundation and were then incorporated into a well-designed report from the Inquiry into British Housing which received extensive publicity (Housing Research Findings no.19 1990; Maclennan et al. 1990). Work with the news media, a public launch at a major conference, presentations to government ministers, all enhanced the research. Ten years later, the Inquiry’s secretary could claim a number of policy changes – including the abolition of mortgage interest tax relief – that were initiated or assisted by this research programme (Best 2002).

Policy and Academia: an Assessment

279

Quality and timeliness Although it might be hoped that the quality of the research would be of central importance in determining its impact, brilliant work sometimes remains unacknowledged and second-rate research sometimes proves highly inﬂuential. Undoubtedly all researchers will strive for quality (and this chapter is not the place to deﬁne it). But topicality and timeliness may prove more potent in the mix when it comes to determining the impact. Funders can be part of the process of looking ahead to assess the likely topicality and wider interest that will greet a research report. Certainly, it can be galling for us to publish useful analysis just after new legislation is enacted on that same subject – with little prospect of further change for many years thereafter. But work done in parallel with a consultative process, perhaps emerging after a Green Paper that seeks an input to the formulation of new policy, can be disproportionately valuable. Thus, a report from Duncan Maclennan on neighbourhood renewal in 2000 was published to coincide with the ﬁnal stages of the government’s preparation of a National Strategy for Neighbourhood Renewal, with presentations accordingly to the key civil servants in the Cabinet Ofﬁce (Maclennan 2000). Seldom will timing be perfect. Unpredictable events in the wider world are sure to intervene. For example, the Joseph Rowntree Foundation’s Income and Wealth Inquiry report, demonstrating how the gap between richer and poorer had widened dramatically in the 1980s, came out at an important moment in the electoral cycle and attracted extensive media interest (Sir Peter Barclay 1995; Hills 1995). Some credit for its timeliness might be attributable to the JRF’s forward planning. But much more important to getting substantial coverage was the coincidence of national revulsion in the press over the high salaries which ‘fat cats’ in the newly-privatised utilities were awarding themselves. It was luck as much as good judgement that meant this major programme (based on research undertaken over several years in over twenty separate academic institutions) caught the public imagination. But even if luck is sure to play a part, it remains important to consider the timeliness of the research before it begins and to choose the moment for its dissemination. Sometimes the window of opportunity will be much longer than in other cases: work on the outcomes for children of divorce and separation and of the relative success of different ways of moderating the impact of negative effects, is likely to be almost as valid whenever it emerges (for example, Rodgers & Pryor 1998). But research into the effectiveness of, say, policy for rent restructuring in social housing will need to conclude in time to meet a government deadline to review that policy. Or insights and evidence that

280

Housing Economics and Public Policy

could inﬂuence public spending decisions may be necessary before a government concludes one of its comprehensive spending reviews.

Persistence and partnership Whether research missed the boat ﬁrst time round, or whether it was simply ahead of its time, it often requires its messages to be repeated – sometimes over several years – before the penny drops or before the moment is right. But the researcher may well have moved on; funds have been available for other priorities and not for following through and repeatedly revisiting earlier work. While this characteristic of the research process highlights the importance of getting the timing right ﬁrst time, it also highlights the theme of the rest of this chapter: this is that researchers may need partners. It may not be sensible to expect research outputs to translate into policy change on their own. Rather it may be wise to conceive of research as but a part of a broader process in the development of policy and practice. Academics keen to see their work used to real purpose may ﬁnd that a successful outcome requires partnership with an agency for which policy change is a fundamental goal. Partner organisations can bring to the table a capacity for persistence. Staying with the subject of the research project long after the work is done presents the chance to draw down the evidence at the moment when the work has become most relevant and topical. The partner organisation can pull the pieces together from research evidence accumulated over some years and deliver it when the policy-making world is ready. (This process is being helped by the accessibility of a back catalogue of related research projects on the websites of intermediary bodies of this kind. See, for example, the Joseph Rowntree Foundation’s website, www.jrf.org.uk, which holds summaries of nearly 1000 reports including a large number by Duncan Maclennan: Maclennan 1993a, 1994; Maclennan & Meen 1993; Maclennan et al. 1991b, c, d.) However, intermediary agencies dedicated to policy change are thin on the ground. Some – like the Policy Studies Institute and the Family Policy Studies Centre – have been unable to sustain their independent existence through shortages of funding. The weakness or absence of these partners makes it harder for the research community to turn academic work into policy change. In the absence of such agencies able to enhance presentation and dissemination, and to ensure the timeliness of the intervention in the policy arena and

Policy and Academia: an Assessment

281

to be persistent in this quest, could academia itself rise to the challenge? Might there be ways in which universities – possibly through specialist dedicated units – could acquire the necessary skills in-house, obtain the required ﬁnancial resources, and act as the engine for using valuable research to make its impact on social policy?

In conclusion It may be neither fair nor realistic to expect the research community’s academic expertise, analytical skills and intelligence not only to provide high quality research outputs but also to be proﬁcient in the techniques of policy development. Funding for research – often short-term, project-by-project with little capacity for taking a continuing line over a sustained period – does not help. And the mechanisms for rewarding researchers – bestowing prestige and promotion on the basis of the impression made, not least through academic journals, on fellow academics and not on those in the world outside – may be inimical to academia being inﬂuential in policy circles (other than through individuals making the transition on a personal basis in the ways I have discussed). It might then be argued that only by the use of intermediary bodies that can add the extra expertise and resources to make research relevant, accessible and topical, will the output of academia be likely to achieve its fullest potential in securing policy change. But it is hard to ﬁnd intermediary agencies that have the capability of acting as supportive partners of this kind in the social policy ﬁeld. Could it be that the universities themselves might be the best places for such complementary skills to be acquired, alongside colleagues who can deliver research of the highest quality? I throw back this challenge to academia – and perhaps to Professor Duncan Maclennan who has shown us all how inﬂuential the skills of the academic can be in supporting policy change at the highest levels.

Notes (1) Currently, over £8 million per annum. (2) Webb resigned from his academic post in 1997 after the general election, but was subsequently made a Visiting Professor at the University. (3) The Joseph Rowntree Foundation publishes summaries of research reports using a standard format known as FINDINGS. This approach has been copied by a large number of organisations (including many overseas). The discipline of compressing the contents of a lengthy document onto four sides of A4 can

282

Housing Economics and Public Policy also concentrate the author’s mind on what exactly has been discovered. This technique – perfected by the Foundation’s Director of Communications, the late Roland Hirst, in 1990 – can help the busy policy maker or practitioner to get to grips with research outputs. It may also stimulate an interest in reading a longer text. Accessibility and readability – preferably with a length in the region of 15 000 words rather than 150 000 words – remains critical for the main report too. At the JRF we would add that the way the document itself is presented will also make a difference; the layout and print style do matter, preferably with the use of colour and an attractive cover (and with a synopsis of the contents on the back).

References

Abraham, J.M. & Hendershott, P.H. (1996) Bubbles in metropolitan housing markets. Journal of Housing Research, 7, 191–207. Adair, A.S., Berry, J. & McGreal, W.S. (1996) Hedonic modelling, housing submarkets and residential valuation. Journal of Property Research, 13 (1), 67–84. Adair, A., McGreal, S., Smyth, A., Cooper, J. & Ryley, T. (2000) House prices and accessibility: The testing of relationships within the Belfast urban area. Housing Studies, 15 (5), 699–716. Adams, D. & May, H. (1992) The role of landowners in the preparation of statutory local plans. Town Planning Review, 63 (3), 297–321. Adams, D., Disberry, A., Hutchinson, N. & Munjoma, T. (1999) Do Landowners Constrain Urban Redevelopment? Aberdeen papers in Land Economy. University of Aberdeen, Dept. of Land Economy, Aberdeen. Akerlof, G. (1980) A theory of social custom, of which unemployment may be one consequence. Quarterly Journal of Economics, 94, 749–75. Allen, F. & Gale, D. (2000) Comparing Financial Systems. The MIT Press, Cambridge, Mass. Allen, M.T., Springer, T.M. & Waller, N.G. (1995) Implicit pricing across residential submarkets. Journal of Real Estate and Financial Economics, 11 (2), 137–51. Alonso, W. (1964) Location and Land Use: Toward a General Theory of Land Rent. Harvard University Press, Cambridge, Mass. Ambrose, B.W. & Nourse, H.O. (1993) Factors inﬂuencing capitalization rates. Journal of Real Estate Research, 8 (2), 221–237. Amemiya, T. (1980) Selection of regressors. International Economic Review, 21, 331–54. Anas, A. & Arnott, R. (1991) Dynamic housing market equilibrium with taste heterogeneity, idiosyncratic perfect foresight and stock conversions. Journal of Housing Economics, 1, 2–32. Anas, A. & Arnott, R. (1993a) Technological progress in a model of the housing-land cycle. Journal of Urban Economics, 34, 186–206. Anas, A. & Arnott, R. (1993b) A fall in construction costs can raise housing rents. Economic Letters, 41, 221–24.

284

References

Anas, A. & Arnott, R. (1993c) Development and testing of the Chicago prototype housing market model. Journal of Housing Research, 4, 73–129. Anas, A. & Arnott, R. (1994) The Chicago prototype housing market model with tenure choice and its policy implications. Journal of Housing Research, 5, 23–90. Anas, A. & Arnott, R. (1997) Taxes and allowances in a dynamic equilibrium model of urban housing with a size-quality hierarchy. Regional Science and Urban Economics, 27, 547–80. Anas, A. & Eum, S.J. (1984) Hedonic analysis of a housing market in disequilibrium. Journal of Urban Economics, 15, 87–106. Anas, A., Arnott, R. & Small, K.A. (1998) Urban spatial structure. Journal of Economic Literature, XXXVI, 1426–64. Anderson, I., Kemp, P.A. & Quilgars, D. (1993) Single Homeless People. HMSO, London. Angel, S. (2000) Housing Policy Matters: A Global Analysis. Oxford University Press. Oxford. Angel, S., Mayo, S.K. & Stephens, W. (1993) The housing indicators program: A report on progress and plans for the future. Netherlands Journal of Housing and the Built Environment, 8 (1), 13–47. Anglin, P.M. & Ramazan, G. (1996) Semiparametric estimation of a hedonic price function. Journal of Applied Econometrics, 11 (6), 633–48. Annez, P. & Wheaton, W.C. (1984) Economic development and the housing sector: a cross-national model. Economic Development and Cultural Change, 32 (4), 749–66. Anstie, R., Findlay, C. & Harper, I. (1983) The impact of inﬂation and taxation on tenure choice and the redistribution effects of home-mortgage interest rate regulation. The Economic Record, June, 105–110. Arnott, R. (1987) Economic Theory and Housing. In: Handbook of Regional and Urban Economics (ed. E. Mills), Vol. 2, chap. 24. North Holland, Amsterdam. Arnott, R., Davidson, R. & Pines, D. (1983) Housing quality, maintenance and rehabilitation. Review of Economic Studies, 50, 467–94. Arrow, K. (1971) Essays in the Theory of Risk Bearing. North Holland, Amsterdam. Arrow, K. J. & Lind, R.C. (1970) Uncertainty and the evaluation of public investments. American Economic Review, June, 364–378. Artle, R. & Varaiya, P. (1978) Life Cycle Consumption and Homeownership. Journal of Economic Theory, 18, 35–58. Atkinson, A.B. & King, M.A. (1980) Housing policy, taxation and reform. Midland Bank Review, Spring, 7–16. Attanasio, O.P. & Weber, G. (1994) The U.K. consumption boom of the late 1980s: Aggregate implications of microeconomic evidence. Economic Journal, 104, 1269–1302. Audit Commission (1992) Developing Local Authority Housing Strategies. HMSO, London. Auten, G.E. & Clotfelter, C. (1982) Permanent vs. transitory tax effects and the realisation of capital gains. Quarterly Journal of Economics, 97, 613–632.

References

285

Awan, K., Odling-Smee, J. & Whitehead, C.M.E. (1982) Household attributes and the demand for private rental housing. Economica, 49, 183–200. Bailey, M.J., Muth, R.F. & Nourse, H.O. (1963) A regression method for real estate price index construction. Journal of the American Statistical Association, 58, 933–42. Bajic, V. (1985) Housing-market segmentation and demand for housing attributes: some empirical ﬁndings. Journal of American Real Estate and Urban Economics Association, 13, 58–75. Baker, R., Challen, P., Maclennan, D., Reid, V. & Whitehead, C.M.E. (1992) The Scope for Competitive Tendering of Housing Management. HMSO, London. Ball, M. (1973) Recent empirical work on the determinants of relative house prices. Urban Studies, 10, 213–33. Ball, M. (1996a) Housing and Construction: a troubled relationship? Policy Press, Bristol. Ball, M. (1996b) Investing in New Housing: lessons for the future. Policy Press, Bristol. Ball, M. & Kirwan, R. (1977) Accessibility and supply constraints in urban housing markets. Urban Studies,14, 11–32. Ball, M. & Wood, A. (1999) Housing investment: Long run international trends and volatility. Housing Studies, 14, 185–210. Ballas, D. & Clarke, G. (1999) Regional Versus Local Multipliers of Economic Change? A micro-simulation approach. Paper presented to the 39th European Regional Science Association Congress, University College Dublin, Dublin, Ireland, 23–27 August. Bannister, J., Dell, M., Donnison, D., Fitzpatrick, S. & Taylor, R. (1993) Homeless Young People in Scotland. HMSO, Edinburgh. Barclay, Sir Peter (chair) (1995) Income and Wealth. Vol. 1 of Report of the Inquiry Group, Joseph Rowntree Foundation, York. Barkham, R. (1992) Regional variations in entrepreneurship: some evidence from the United Kingdom. Entrepreneurship & Regional Development, 4, 225–244. Barlow, J. (1993) Controlling the housing land market: some examples from Europe. Urban Studies, 30 (7), 1129–50. Barlow, J., Cocks, R. & Parker, M. (1994) Planning for Affordable Housing. HMSO, London. Barr, N. (1998) The Economics of the Welfare State. Oxford University Press, Oxford. Bartik, T. J., Butler, J.S. & Liu, J. (1992) Maximum score estimates of the determinants of residential mobility: implications for the value of residential attachment and neighborhood amenities. Journal of Urban Economics, 32 (2), 233–56. Bartlett, W. (1991) Housing Supply Elasticities: a review. Housing Finance, York. Bassett, K. (1980) Council house sales in Bristol 1960–79. Policy and Politics, 8, 324–33. Basu, S. & Thibodeau, T.G. (1998) Analysis of spatial autocorrelation in house prices. Journal of Real Estate Finance and Economics, 17 (1), 61–85. Bate, R. (1999) A Guide to Land Use and Housing. Overview Paper for Joseph Rowntree Foundation, RESC Programme.

286

References

Beazley, M. (1980) The sale of council houses in a rural area: a case study of South Oxfordshire. Working paper 44, Oxford Polytechnic Department of Town Planning, Oxford. Berg, L. (1994) Household savings and debts: The recent experience of the Nordic countries. Oxford Review of Economic Policy, 10, 42–53. Berg, L. & Bergstrom, R. (1995) Housing and ﬁnancial wealth, ﬁnancial deregulation and consumption – the Swedish case. Scandinavian Journal of Economics, 97, 421–39. Berkovec, J.A., Canner, G.B., Gabriel, S.A., & Hannan, T.H. (November 1994) Discrimination, default, and loss in FHA mortgage lending. Federal Reserve Board of Governors Working Paper. Berry, B.J.L. & Bednarz, R. (1975) A hedonic model of prices and assessments for single family homes in Chicago: Does the assessor follow the market or the market follow the assessor? Land Economics, 51, 21–40. Berry, B.J.L., Chung, K.S., & Waddell, P. (1995) Widening gaps: the behaviour of submarket housing prices indexes in the Dallas area, 1979–1993. Urban Geography, 16 (8), 722–33. Best, R. (2002) Inquiry into British Housing 1984–1991: What has happened since. Joseph Rowntree Foundation, York. Black, J., de Meza, D. & Jeffreys, D. (1996) House prices, the supply of collateral and the enterprise economy. Economic Journal, 106, 60–75. Blackaby, D.H. & Manning, D.N. (1992) Regional earnings and unemployment – A simultaneous approach. Oxford Bulletin of Economics and Statistics, 54, 481–502. Blackaby, R. (2000) Understanding Local Housing Markets: Their role in local housing strategies. Chartered Institute of Housing/Council of Mortgage Lenders, Coventry and London. Blomquist, G. & Worley, L. (1982) Specifying the demand for housing characteristics: The exogeneity issue. In: The Economics of Urban Amenities (ed. D.B. Diamond & G. Tolley). Academic Press, London. Boelhouwer, P.J. (2000) Convergence in European mortgage systems before and after EMU. Journal of Housing and the Built Environment, 15, 11–28. Boelhouwer, P. & Van der Heijden, H. (1993) Housing systems in Europe – A research project. Scandinavian Housing Planning Research, 10 (1), 37–42. Bös, D. (1983) Public pricing with distributional objectives. Public Sector Economics (ed. J. Finsinger). Macmillan, London. Bös, D. (1985) Public sector pricing. In: Handbook of Public Economics (ed. A.J. Anerbach & M. Feldstein). North Holland, Amsterdam. Bossons, J. (1978) Housing demand and household wealth: Evidence for homeowners. In: Urban Housing Markets: Recent Directions in Research and Policy (ed. L. B. Bourne and J. R. Hitchcock). University of Toronto Press, Toronto. Bourassa, S.C. & Grigsby, W.G. (2000) Income tax concessions for owner-occupied housing. Housing Policy Debate, 11 (3), 521–46. Bourassa, S.C. & Hendershott, P.H. (1994) On the equity effects of taxing imputed rent: Evidence from Australia. Housing Policy Debate, 5 (1), 73–95.

References

287

Bourassa, S.C., Hamelink, F., Hoesli, M. & MacGregor, B.D. (1999) Deﬁning housing submarkets. Journal of Housing Economics, 8, 160–183. Bover, O., Muellbauer, J. & Murphy, A. (1989) Housing, wages and U.K. labour markets. Oxford Bulletin of Economics and Statistics, 51, 97–136. Bowden, R.J. (1978) The Econometrics of Disequilibrium. North Holland, Amsterdam. Box, G.E.P. & Cox, D. (1964) An analysis of transformations. Journal of the American Statistical Association, Society Series B, 26, 211–52. Boyle, M.A. & Kiel, K.A. (2001) A survey of house price hedonic studies of the impact of environmental externalities. Journal of Real Estate Literature, 9 (2), 117–44. Bramley, G. (1989) Land supply, planning and private housebuilding: a review. SAUS Working Paper 81. School for Advanced Urban Studies, Bristol. Bramley, G. (1993a) The impact of land use planning and tax subsidies on the supply and price of housing in Britain. Urban Studies, 30, 5–30. Bramley, G. (1993b) Land use planning and the housing market in Britain: the impact on housebuilding and house prices. Environment and Planning A, 25, 1021–51. Bramley, G. (1998) Measuring planning: indicators of planning restraint and its impact on the housing market. Environment and Planning B: Planning & Design, 25, 31–57. Bramley, G. (1999) Housing market adjustment and land-supply constraints. Environment and Planning A, 31, 1169–88. Bramley, G. & Lancaster, S. (1998) Household formation: A suitable case for policy? Housing Finance, 38, 20–29. Bramley, G. & Smart, G. (1996) Modelling local income distributions in Britain. Regional Studies, 30, (3), 239–55. Bramley, G. & Watkins, C. (1996a) Steering the Housing Market: new building and the changing planning system. The Policy Press, Bristol. Bramley, G. & Watkins, C. (1996b) Modelling the relationship between land availability, the land-use planning system and the supply of new housing. Paper presented at RICS ‘Cutting Edge’ Conference, University of the West of England, Bristol, 20–21 Sept. Bramley, G., Bartlett, W., & Lambert, C. (1995) Planning, the Market and Private Housebuilding. UCL Press, London. Bramley, G., Munro, M. & Lancaster, S. (1997) Economic Inﬂuences on Household Formation: A Literature Review. Department of Environment, Transport and the Regions, London. Bramley, G., Pawson, H., Satsangi, M. & Third, H. (1999) Local housing needs assessment: a review of current practice and the need for guidance. Research Paper No. 73, School of Planning and Housing, Edinburgh College of Art/HeriotWatt University, Edinburgh. Bramley, G., Pawson, H. & Parker, J. (2000) Local housing needs assessment: a guide to good practice. Department of Environment, Transport and the Regions, London. Bratt, R.G. (2002) Housing and Family Well-being. Housing Studies, 17 (1), 13–26. Breheny, M. (1999) People, households and houses: The basis to the ‘great housing debate’ in England. Town Planning Review, 70, 275–93.

288

References

Britton, E. & Whitley, J. (1997) Comparing the monetary transmission mechanism in France, Germany and the United Kingdom: Some issues and results. Bank of England Quarterly Bulletin, May, 152–62. Brock, W. & Durlauf, S. (forthcoming) Interactions-based models. In: Handbook of Econometrics (ed. J. Heckman & E. Learner), vol. 5. North Holland, Amsterdam. Bromwich, M., Harrison, A. Travers, A. & Whitehead, C.M.E. (2002) An economic analysis of PFI. LSE London Discussion Paper, forthcoming, LSE, London. Brown, J., Song, H. & McGillvray, A. (1997) Forecasting U.K. house prices: A time varying coefﬁcient approach. Economic Modelling, 14, 529–48. Brueckner, J. K. (1986) The down payment constraint and housing tenure choice: A simpliﬁed exposition. Regional Science and Urban Economics, 16 (4), 519–25. Brueggemann, W.B., Fisher, J.D. & Stern, J.J. (1982) Rental housing and the Economic Recovery Tax Act of 1981. Public Finance Quarterly, 10 (2), 222–41. Buchel, S. & Hoesli, M. (1995) A hedonic analysis of rent and rental revenue in the subsidised and unsubsidised housing sectors in Geneva. Urban Studies, 32 (7), 1199–1213. Burgess, J.F. Jr. & Harmon, O.R. (1982) Speciﬁcation tests in hedonic models. Journal of Real Estate Finance and Economics, 4 (4), 375–93. Burns, L. & Grebler, L. (1977) The Housing of Nations. Macmillan, London. Burrows, R., Pleace, N. & Quilgar, D. (1997) Homelessness and Social Policy. Routledge, London. Business Strategies Limited (1998) A Critique of the Ofﬁcial Population and Household Projections for Scotland. Business Strategies Limited, London. Butler, R.V. (1982) The speciﬁcation of hedonic indexes for urban housing. Land Economics, 58 (1), 96–108. Cameron, G. & Muellbauer, J. (2000) Earnings, unemployment and housing: evidence from a panel of British regions. CEPR Discussion Paper No. 2404. Can, A. (1992) Speciﬁcation and estimation of hedonic housing price models. Regional Science and Urban Economics, 22(3), 453–75. Capozza, D. & Seguin, P.J. (1995) Expectations, efﬁciency and euphoria in the housing market. National Bureau of Economic Research, Working Paper No. 5179. Capozza, D.R., Green, R.K. & Hendershott, P.H. (1996) Taxes, mortgage borrowing and residential land prices. In: Economic Effects of Fundamental Tax Reform (ed. H. Aaron & W. Gale). The Brookings Institute, Washington, D.C. Capozza, D.R., Green, R.K. & Hendershott, P.H. (1997) Income taxes and house prices. CULER Working Paper. Carr-Hill, R. (1997) Impact of housing conditions upon health status. Paper presented at seminar on The Wider Impacts of Housing, Scottish Homes. Case, B. & Quigley, J.M. (1991) The dynamics of real estate prices. Review of Economics and Statistics, 22 (1), 50–8. Case, K.E. & Shiller, R.J. (1987) Prices of single family homes since 1970: New indexes for four cities. New England Economic Review, September/October, 45–56. Case, K.E. & Shiller, R.J. (1989) The efﬁciency of the market for single family homes. American Economic Review, 79, 125–37.

References

289

Case, K.E. & Shiller, R.J. (1990) Forecasting prices and excess returns in the housing market. AREUEA Journal, 18, 253–73. Case, K.E., Quigley, J.M. & Shiller, R.J. (2001) Stock market wealth, housing wealth, spending and consumption. Unpublished. Cassel, E. & Mendelsohn, R. (1985) The choice of functional forms for hedonic price equations: comment. Journal of Urban Economics, 18, 135–42. Centre for Housing Research (1989) The Nature and Effectiveness of Housing Management in England. HMSO, London. Chambers, D. (1992) The racial housing price differential and racially transitional neighborhoods. Journal of Urban Economics, 32(2), 214–32. Chambers, D. & Simonson, J. (1989) Mobility, transactions costs, and tenure choice. Paper presented at the Seventeenth Midyear Meeting of The American Real Estate and Urban Economics Association, May. Champion, A.J., Fotheringham, S., Boyle, P., Rees, P., & Stilwell, J. (1998b) The determinants of migration ﬂows in England: a review of existing data and evidence. Report to DETR. DETR, London. Champion, A.J., Coombes, M., Fotheringham, S. et al. (2000) Development of a migration model. Final report, prepared for the Department of the Environment, Transport and the Regions (DETR), under Contract RADS 5/9/22, University of Newcastle upon Tyne, Newcastle. Chan, S. (2001) Spatial lock-in: Do falling house prices constrain residential mobility? Journal of Urban Economics, 49, 567–86. Chaplin, R., Jones, M., Martin, S. et al. (1995) Rents and Risks: Investing in Housing Associations. Joseph Rowntree Foundation, York. Charles, K. & Hurst, E. (forthcoming) The Transition to Home Ownership and the Black-White Wealth Gap. Review of Economics and Statistics. Charles, S. (1977) Housing Economics. Macmillan, London. Cheshire, P. & Leven, C. (1982) On the costs and economic consequences of the British land use planning system. Discussion paper in Urban and Regional Economics, Series C, no. 11, University of Reading. Cheshire, P. & Sheppard, S. (1989) British planning policy and access to housing: some empirical estimates. Urban Studies, 26, 469–85. Cheshire, P. & Sheppard, S. (1995) On the price of land and the value of amenity. Econometrica, 62, 247–67. Cheshire, P. & Sheppard, S. (1997) The welfare economics of land use regulation. Research Papers in Environmental and Spatial Analysis No. 42. Department of Geography, London School of Economics. Chinloy, P. (1991) Risk and the user cost of housing services. AREUEA Journal, 19 (4), 516–31. Cho, M. (1996) House price dynamics: A survey of theoretical and empirical issues. Journal of Housing Research, 7, 145–72. Chowhan, J. & Prud’homme, M. (2000) City Comparisons of Shelter Costs in Canada: A Hedonic Approach. Statistics Canada, Ottawa. Christensen, L.R., Jorgenson, D.W. & Lau, L.J. (1973) Transcendental logarithmic production frontiers. Review of Economics and Statistics, 55, 28–45.

290

References

Clapham, D. (1996) Housing and the economy: Broadening comparative housing research. Urban Studies, 33, 631–48. Clapham, D. & Maclennan, D. (1983) Residualization of council housing: a nonissue. Housing Review, 32, 9–10. Clapp, J.M. & Giaccotto, C. (1998a) Price indices based on the hedonic repeat-sale method: Application to the housing market. Journal of Real Estate Finance and Economics, 16 (1), 5–26. Clapp, J.M. & Giaccotto, C. (1998b) Residential hedonic models: A rational expectations approach to age effects. Journal of Urban Economics, 44 (3), 415– 37. Clapp, J.M., Giaccotto, C. & Tirtiroglu, D. (1991) Housing price indices: Based on all transactions compared to repeat subsamples. AREUEA Journal, 19 (3), 270–85. Clark, W.A.V. (ed.) (1982a) Modelling housing market search. Croom Helm, London. Clark, W.A.V. (1982b) Recent research on migration and mobility. Progress and Planning, 18, 1–56. Clark, W.A.V. (1993) Search and choice in urban housing markets. In: Behavior and Environment: Psychological and Geographical Approaches (eds T. Gärling & R. G. Golledge). Elsevier, Amsterdam. Clark, Todd E. (1995) Rents and prices of housing across areas of the United States: A cross-section examination of the present value model. Regional Science and Urban Economics, 25, 237–47. Cliff, A.D., Haggett, P., Ord, J.K., Bassett, K.A. & Davies, R.B. (1975) Elements of spatial structure. Cambridge University Press, Cambridge, England. Colwell, P.F. & Dilmore, G. (1999) Who was ﬁrst? An examination of an early hedonic study. Land Economics, 75 (4), 620–6. Coulson, N.E. & Fisher, L.M. (2002) Tenure choice and labour market outcomes. Housing Studies, 17 (1), 35–49. Council for the Protection of Rural England (2001) Sprawl Patrol: ﬁrst year report. CPRE, London. Court, A.T. (1939) Hedonic price indexes with automotive examples. The Dynamics of Automobile Demand. General Motors Corporation, New York. Crane, J. (1991) The epidemic theory of ghettos and neighbourhood effects on dropping out and teenage childbearing. American Journal of Sociology, 96, 1226–59. Cressy, R. (1996) Are business startups debt-rationed? Economic Journal, 106, 1253–70. Cullingworth, J.B. (1997a) Planning in the U.S.A.: Policies, Issues and Processes. Routledge, London & New York. Cullingworth, J.B. (1997b) British land-use planning: A failure to cope with change? Urban Studies, 34, 945–60. Cunningham, D.F. & Hendershott, P.H. (1984) Pricing FHA Mortgage Default Insurance. Housing Finance Review, 3, 373–92. Dale-Johnston, D. (1982) An alternative approach to housing market segmentation using hedonic price data. Journal of Urban Economics, 11, 311–32. Davies, H.W.E. (1998) Continuity and change: the evolution of the British Planning System, 1947–97. Town Planning Review, 69, 135–52.

References

291

Davis, O.A. & Whinston, A.B. (1961) The economics of urban renewal. Law and Contemporary Problems, 26 (1), 163–77. De Borger, B. (1986) Estimating the benefits of public housing programs: A characteristics approach. Journal of Regional Science, 26, 761–73. De Leeuw, F. & Ozanne, L. (1981) Housing. In: How Taxes Affect Economic Behaviour (eds H. J. Aaron & J. A. Pechman). Brookings Institute, Washington, DC. De Leeuw, F. & Struyck, R. (1975) The Web of Urban Housing. The Urban Institute, Washington, DC. Department of Environment (1971) Fair Deal for Housing, Cmnd 4728, HMSO, London. Department of the Environment (1977) Housing Policy: a consultative document. Cmnd 6851, HMSO, London. Department of the Environment (1980) Housing Requirements: a guide to information and techniques. HMSO, London. Department of Environment, Transport and the Regions (1997) An Economic Model of the Need and Demand for Social Housing. DETR, HMSO, London. Department of Environment, Transport and the Regions (1998) The 1996 House Condition Survey. HMSO, London. Department of Environment, Transport and the Regions (1999a) Projections of Households in England to 2021. DETR, HMSO, London. Department of Environment, Transport and the Regions (1999b) Towards an urban renaissance. Final report of the Urban Task Force under the chairmanship of Lord Rogers of Riverside. HMSO, London. Department of Transport, Local Government and the Regions (2001) Reforming planning obligations: delivering fundamental change. Consultation paper. Des Rosiers, F. & Theriault, M. (1996) Rental amenities and the stability of hedonic prices: A comparative analysis of ﬁve market segments. Journal of Real Estate Research, 12 (1), 17–36. Diamond, D.B. Jr. & Smith, B. (1985) Simultaneity in the Market for Housing Characteristics. Journal of Urban Economics, 17, 280–92. DiPasquale, D. & Glaeser, E. (1999) Incentives and social capital: Are homeowners better citizens? Journal of Urban Economics, 45 (2), 354–84. DiPasquale, D. & Wheaton, W. (1994) Housing market dynamics and the future of house prices. Journal of Urban Economics, 35, 1–27. DiPasquale, D. & Wheaton, W. (1996) Urban Economics and Real Estate Markets. Prentice Hall, Englewood Cliffs, NJ. Din, A., Hoesli, M. & Bender, A. (2001) Environmental values and real estate prices. Urban Studies, 38 (100), 1989–2000. Dobson, J. (2000) Strategic approaches. Joseph Rowntree Foundation, Search No 34, winter 2000/2001, 8–11. Donnison, D. (1961) The teaching of social administration. British Journal of Sociology, 12 (3), 202–23. Donnison, D. (2000) The Academic Contribution to Social Reform. Social Policy and Administration, 34 (1), 26–43. Doyal, L. & Gough, I. (1991) A Theory of Human Need. Macmillan, London.

292

References

Dreiman, M. & Follain, J.R. (2000) Drawing Inferences about Housing Supply Elasticity from House Price Responses to Income Shocks. (Unpublished). Dubin, R.A. (1992) Spatial autocorrelation and neighborhood quality. Regional Science and Urban Economics, 22 (3), 433–52. Dubin, R.A. (1998) Predicting house prices using multiple listings data. Journal of Real Estate Finance and Economics, 17 (1), 35–59. Duca, J.V. & Rosenthal, S.S. (1994) Borrowing constraints and access to owneroccupied housing. Regional Science and Urban Economics, 24, 301–22. Dynarski, M. (1986) Residential attachment and housing demand. Urban Studies, 23, 11–20. Dynarski, M. & Sheffrin, S.M. (1985) Housing purchases and transitory income: A study with panel data. The Review of Economics and Statistics, 67, 195–205. Edel, M. & Sclar, E. (1974) Taxes, spending and property values: Supply adjustment in a Tiebout-Oates Model. Journal of Political Economy, 82, 941–54. Engelhardt, G.V. (1994a) Tax subsidies to saving for home purchase: Evidence from Canadian RHOSPs. National Tax Journal, 47 (2), 363–88. Engelhardt, G.V. (1994b) House prices and the decision to save for down payments. Journal of Urban Economics, 36 (2), 209–37. Engelhardt, G.V. (1995) House prices and home owner saving behavior. National Bureau of Economic Research Working Paper, no. 5183. Engelhardt, G.V. & Mayer, C.J. (1994) Gifts for home purchase and housing market behavior. New England Economic Review (May-June), 47–58. Engelhardt, G.V. & Mayer, C.J. (1995) Intergenerational transfers, borrowing constraints, and saving behavior: Evidence from the housing market. Federal Reserve Bank of Boston Working Paper, no. 95–11. Englund, P. & Ioannides, Y. (1997) House price dynamics: An international empirical perspective. Journal of Housing Economics, 6, 119–36. Englund, P., Quigley, J.M. & Redfearn, C.L. (1998) Improved price indexes for real estate: Measuring the course of Swedish housing prices. Journal of Urban Economics, 44 (2), 171–96. Environment Committee (1981) Council house sales. Second Report HC 366-I, HC 535 I-xi (1979–80), HMSO, London. Epple, D. (1987) Hedonic prices and implicit markets: Estimating the demand and supply functions for differentiated products. Journal of Political Economy, 95, 59–80. Ermisch, J. (1984) Housing Finance: Who Gains? Policy Studies Institute, London. Ermisch, J.F., Findlay, J. & Gibb, K. (1996) The price elasticity of housing demand in Britain: Issues of sample selection. Journal of Housing Economics, 5 (1), 64–86. Evans, A. (1973) Economics of Residential Location. Macmillan, London. Evans, A.W. (1983) The determination of the price of land. Urban Studies, 20, 119–29. Evans, A.W. (1991) Rabbit hutches on postage stamps: planning, development and political economy. Urban Studies, 28 (6), 853–70. Evans, A.W. (1996) The impact of land use planning and tax subsidies on the supply and price of housing in Britain. Urban Studies, 33, 581–6.

References

293

Evans, A., Stevens, S. & Williams, P. (1999) Making Best Use of the Private Sector? A review of Welsh local authority housing strategies and operational plans. Council of Mortgage Lenders, London. Fair, R.C. & Jaffee, D.M. (1972) Methods of estimation for markets in disequilibrium. Econometrica, 40, 497–514. Feder, G. (1982) On exports and economic growth. Journal of Development Economics, 12, 59–73. Feldstein, M., Slemrod, J. & Yitzhaki, S. (1980) The effects of taxation on the selling of corporate stock and the realisation of capital gains. Quarterly Journal of Economics, 94 (4), 777–91. Fischel, W. (1990) Do Growth Controls Matter? A review of empirical evidence on the effectiveness and efﬁciency of local government land use regulation. Lincoln Institute of Land Policy Working Paper, Lincoln Institute, Boston. Fischer, C.S. (1982) To Dwell Among Friends. University of Chicago Press, Chicago. Fisher, F. & Shell, K. (1971) Taste and quality change in the pure theory of the true cost of living index. In: Price Indexes and Quality Change (ed. Zvi Griliches), Harvard University Press. Fisher, J.D. & Lentz, G.H. (1986) Tax reform and the value of real estate income property. AREUEA Journal, 14 (2), 287–315. Flood, J. (1997) Urban and housing indicators. Urban Studies, 34 (10), 1597–1620. Follain, J.R. (1982) Does inﬂation affect real behavior? The Case of Housing. Southern Economic Journal, 48 (3), 570–82. Follain, J.R. & Jimenez, E. (1985a) Estimating the demand for housing characteristics: A survey and critique. Regional Science and Urban Economics, 15 (1), 77–107. Follain, J.R. & Jimenez, E. (1985b) The demand for housing characteristics in developing countries. Urban Studies, 22 (5), 421–32. Follain, J.R. & Ling, D.C. (1988) Another look at tenure choice, inﬂation, and taxes. Journal of the American Real Estate and Urban Economics Association, 16 (3), 207–29. Follain, J.R. & Malpezzi, S. (1980a) Dissecting Housing Value and Rent. The Urban Institute, Washington, D.C. Follain, J.R. & Malpezzi, S. (1980b) Estimates of housing inﬂation for thirty-nine SMSAs: An alternative to the consumer price index. Annals of Regional Science, 41–56. Follain, J.R. & Malpezzi, S. (1981a) Another look at racial differences in housing prices. Urban Studies, 18 (2), 195–203. Follain, J.R. & Malpezzi, S. (1981b) Are occupants accurate appraisers? Review of Public Data Use, 9 (1), 47–55. Follain, J.R. & Malpezzi, S. (1981c) The ﬂight to the suburbs: Insight from an analysis of central city versus suburban housing costs. Journal of Urban Economics, 9 (3), 381–98. Follain, J.R. & Ozanne, L. with Alberger, V. (1979) Place to Place Indexes of the Price of Housing. Urban Institute, Washington, D.C.

294

References

Follain, J.R., Hendershott, P.H. & Ling, D.C. (1987) Understanding the real estate provisions of tax reform: Motivation and impact. National Tax Journal, 3, 363–72. Fordham, R. & Brook, R. (1995) A Methodology for Assessing Housing Need. Planning Negotiators and National Housing and Town Planning Council, London. Forrest, R. & Murie, A. (1976) Social segregation, housing need and the sale of council houses. Research memorandum 53, Centre for Urban and Regional Studies, University of Birmingham. Forrest, R. & Murie, A. (1984a) Right to Buy? Issues of need, equity and polarisation in the sale of council houses. Working Paper 39, School for Advanced Urban Studies, University of Bristol. Forrest, R. & Murie, A. (1984b) Monitoring the Right to Buy 1980–1982. Working Paper 40, School for Advanced Urban Studies, University of Bristol. Forrest, R. & Murie, A. (1988) Selling the Welfare State: the Privatisation of Public Housing. Routledge, London. Forrest, R. & Murie, A. (1990a) Moving the Housing Market. Avebury, Aldershot. Forrest, R. & Murie, A. (1990b) Residualisation and council housing: a statistical update. Working Paper 91, School for Advanced Urban Studies, University of Bristol. Foulis, M.B. (1983) Council House Sales in Scotland. HMSO, Edinburgh. Freeman, A.M. (1979) Hedonic prices, property values and measuring environmental beneﬁts: A survey of the issues. Scandinavian Journal of Economics, 81 (2), 154–73. Friedman, J. & Weinberg, D.H. (1981) The demand for rental housing: Evidence from the housing allowance demand experiment. Journal of Urban Economics, 9 (3), 311–31. Fujita, M. (1987) Urban Economic Theory. Cambridge University Press, Cambridge. Gabriel, S.A. (1984) A note on housing market segmentation in an Israeli development town. Urban Studies, 21, 189–94. Galster, G.C. (1987) Homeowners and Neighborhood Reinvestment. Duke University Press, Durham, NC. Galster, G.C. (1992) Research on discrimination in housing and mortgage markets: Assessment and future directions. Housing Policy Debate, 3 (2), 639–84. Galster, G. (1996) William Grigsby and the analysis of housing submarkets and ﬁltering. Urban Studies, 33 (10), 1797–1806. Galster, G. (1997) Comparing demand-side and supply-side housing policies; submarket and spatial perspectives. Housing Studies, 12 (4), 561–77. Galster, G.C. (forthcoming) The effects of MTO on sending and receiving neighborhoods. In: Choosing a better life? A Social Experiment in Leaving Poverty Behind (eds J. Goering, T. Richardson & J. Feins). Urban Institute Press, Washington, DC. Galster, G.C. & Daniell, J. (1996) Housing. In: Reality and Research: Social Science and U.S. Urban Policy Since 1960 (ed. G.C. Galster), 85–112. Urban Institute Press, Washington, DC.

References

295

Galster, G.C. & Killen, S. (1995) The geography of metropolitan opportunity – a reconnaissance and conceptual framework. Housing Policy Debate, 6 (1), 7–43. Galster, G.C. & Rothenberg, J. (1991) Filtering in urban housing: a graphical analysis of a quality-segmented market. Journal of Planning Education and Research, 11, 37–50. Galster, G.C., Quercia, R.G., & Cortes, A. (2000) Identifying neighborhood thresholds: an empirical exploration. Housing Policy Debate, 11 (3), 701–32. Garrod, G. & Willis, K.G. (1992a) The environmental economic impact of woodland: A two-stage hedonic price model of the amenity value of forestry in Britain. Applied Economics, 24 (7), 715–28. Garrod, G. & Willis, K.G. (1992b) Valuing goods characteristics: An application of the hedonic price method to environmental attributes. Journal of Environmental Management, 34 (1), 59–76. Gatzlaff, D.H. & Haurin, D.R. (1997) Sample selection bias and repeat-sales index estimates. Journal of Real Estate Finance and Economics, 14, 33–50. Gatzlaff, D.H. & Tirtiroglu, D. (1995) Real estate market efﬁciency: Issues and evidence. Journal of Real Estate Literature, 3, 157–92. Gatzlaff, D.H., Green, R.K. & Ling, D.L. (1998) Revisiting cross-tenure differences in housing maintenance. Land Economics, 74 (3), 328–42. Genesove, D. & Mayer, C. J. (2001) Loss aversion and seller behavior: Evidence from the housing market. Quarterly Journal of Economics, 116 (4), 1233–60. Gibb, K. (1989) Housing Economics and the Urban Public Sector. Centre for Housing Research Discussion Paper 26. University of Glasgow. Gibb, K. & MacKay, D. (2001) The demand for housing in Scotland: new estimates from the Scottish House Condition Survey. University of Glasgow, Department of Urban Studies Discussion Paper. Gibb, K., Meen, G. & Mackay, D. (2000) The Demand for Social Rented Housing in Glasgow: Citywide Needs and Demand. Glasgow City Council/Scottish Homes, Glasgow. Gillen, K., Thibodeau, T. & Wachter, S. (2001) Anisotropic autocorrelation in house prices. Journal of Real Estate Finance and Economics, 23 (1), 5–30. Gilley, O.W. & Kelley Pace, R. (1995) Improving hedonic estimation with an inequality restricted estimator. Review of Economics and Statistics, 77 (4), 609–21. Goldberger, A.S. (1968) The interpretation and estimation of Cobb-Douglas functions. Econometrica, 35, 464–72. Goodlad, R. (1993) The Housing Authority as Enabler. Longman/Chartered Institute of Housing, Essex and Coventry. Goodman, A.C. (1981) Housing submarkets within urban areas: deﬁnitions and evidence. Journal of Regional Science, 21, 175–85. Goodman, A.C. (1998) Andrew Court and the invention of hedonic price analysis. Journal of Urban Economics, 44 (2), 291–8. Goodman, A.C. & Thibodeau, T.G. (1995) Age-related heteroskedasticity in hedonic house price equations. Journal of Housing Research, 6 (1), 25–42. Goodman, A.C. & Thibodeau, T.G. (1998) Housing market segmentation. Journal of Housing Economics, 7, 121–43.

296

References

Goodman, J.L. Jr. & Ittner, J.B. (1992) The accuracy of home owners’ estimates of House Value. Journal of Housing Economics, 2(4), 339–57. Gordon, I. (1990) Housing and labour market constraints on migration across the North-South divide. In: Housing and the National Economy (ed. J. Ermisch). Avebury. Gordon, I. (1991) Multi-stream migration modelling. In: Migration models; macro and micro approaches (ed. J. Stillwell & P. Congdon). Belhaven, London. Gordon, R.H., Hines, J.R. (Jr), & Summers, L.H. (1987) Notes on the tax treatment of structures. In: The Effects of Taxation on Capital Accumulation (ed. M. Feldstein). Chicago University Press, Chicago. Goss, S. & Blackaby, B. (1998) Designing Local Housing Strategies. Local Government Association/Chartered Institute of Housing, Coventry and London. Granovetter, M. (1978) Threshold models of collective behavior. American Journal of Sociology, 83, 1420–43. Granovetter, M. & Soong, R. (1986) Threshold models of diversity: Chinese restaurants, residential segregation, and the spiral of silence. Journal of Sociology, 18, 69–104. Grant, M. (1992) Planning law and the British land use planning system. Town Planning Review, 63 (1), 3–12. Grant, R.A., Thomson, B.W., Bible, J.K. & Randall, J.N. (1976) Local Housing Needs and Strategies: A case study of the Dundee sub-region. Scottish Development Department, Edinburgh. Gravelle, J.G. (1994) The Economic Effects of Taxing Capital Income. MIT Press, Cambridge, Mass. Green, R.K. & Hendershott, P.H. (2001a) Homeownership and unemployment in the U.S. Urban Studies, 38 (9), 1509–20. Green, R.K. & Hendershott, P.H. (2001b) Homeownership and the duration of unemployment: a test of the Oswald hypothesis. Paper presented at the AREUEA Annual Meeting, January 2002. Green, R.K. & Malpezzi, S. (2001) A Primer on U.S. Housing Markets and Policies. The Urban Institute Press for the American Real Estate and Urban Economics Association. Green, R.K. & Vandell, K.D. (1999) Giving households credit: How changes in the U.S. Tax Code could promote homeownership. Regional Science and Urban Economics, 29, 419–44. Green, R.K. & White, M.J. (1997) Measuring the beneﬁts of homeowning: Effects on children. Journal of Urban Economics, 41 (3), 441–61. Greer, S. (1962) The Emerging City: Myth and Reality. Free Press, New York. Grey, A., Hepworth, N., & Odling-Smee, J. (1981) Housing rents, costs and subsidies: a discussion document (2nd edn). The Chartered Institute of Public Finance and Accountancy, London. Grigsby, W. (1963) Housing Markets and Public Policy. University of Pennsylvania Press, Philadelphia. Grigsby, W., Baratz, G., Galster, G.C., & Maclennan, D. (1987) The Dynamics of Neighborhood Change and Decline. Pergamon, London, Progress in Planning series no. 28.

References

297

Griliches, Z. (1961) Hedonic price indexes for automobiles: An econometric analysis quality change. In: The price statistics of the Federal Government, General Series, 73, 137–96. NBER, New York. Gross, D.J. (1986) Designing a suitable project: integration of a demand module into a supply side planning model. World Bank, Water Supply and Urban Development Department Discussion Paper No. UDD-103. Grossman, S. & Laroque, G. (1990) Asset pricing and optimal portfolio choice in the presence of illiquid durable consumption goods. Econometrica, 58, 25–51. Guisanni, B. & Hadjimatheou, G. (1991) Modelling regional house prices in the United Kingdom. Papers in Regional Science, 70, 201–19. Gyourko, J. & Linneman, P. (1993) The affordability of the American dream: An examination of the last 30 years. Journal of Housing Research, 4 (1), 39–72. Gyourko, J. & Tracy, J. (1999) A look at real housing prices and incomes: Some implications for housing affordability and quality. Economic Policy Review, 5 (3), 63–77. Gyourko, J., Linneman, P. & Wachter, S. (1999) Analyzing the relationships among race, wealth, and home ownership in America. Journal of Housing Economics, 8 (2), 63–89. Haas, G.C. (1922) Sales prices as a basis for farm land appraisal. Technical Bulletin 9, University of Minnesota Agricultural Experiment Station, St. Paul, Minnesota. Haavio, M. & Kauppi, H. (2001) Housing markets, borrowing constraints and labor mobility. Paper presented at the Sixth Nordic Conference on Local Public Finance, Helsinki, November 2001. Hall, P., Land, H., Parker, R. & Webb, A. (1975) Change, Choice and Conﬂict in Social Policy. Heinemann, London. Hall, S., Psaradakis, Z. & Sola, M. (1997) Switching error-correction models of house prices in the United Kingdom. Economic Modelling, 14, 517–28. Hall, S., Lazarova, S. & Urga, G. (1999) A principal components analysis of common stochastic trends in heterogeneous panel data: Some Monte Carlo evidence. Oxford Bulletin of Economics and Statistics, Special Issue 61, 749–67. Hallman, H.W. (1984) Neighborhoods: Their Place in Urban Life. SAGE Publications, Beverly Hills, California. Halvorsen, R. & Palmquist, R. (1980) The interpretation of dummy variables in semilogarithmic regressions. American Economic Review, 70, 474–5. Halvorsen, R. & Pollakowski, H. (1981) Choice of functional form for hedonic price functions. Journal of Urban Economics, 10, 37–49. Hamilton, B. (1975) Zoning and property taxation in a system of local government. Urban Studies, 12, 205–11. Hancock, K. (1991) The economic principles of affordability. Paper given at Housing Studies Association Conference, York. Hancock, K. & Maclennan, D. (1989) House price monitoring systems and housing planning in Scotland: a feasibility study. Report for the Scottish Ofﬁce, Centre for Housing Research, Glasgow University. Hansmann, H. (1996) The Ownership of Enterprise. Harvard University Press, Cambridge, Mass.

298

References

Harrison, A.J. (1977) Economics and Land Use Planning. Policy Journals, Newbury. Harsman, B. & Quigley, J.M. (1991) Housing Markets and Housing Institutions: An International Comparison. Kluwer, Boston. Haurin, D.R. (1991) Income variability, homeownership and housing demand. Journal of Housing Economics, 1, 60–74. Haurin, D.R., Hendershott, P. & Wachter, S. (1997) Borrowing constraints and the tenure choice of young households. Journal of Housing Research, 8, 137–54. Hausman, J.A. (1978) Specification tests in econometrics. Econometrica, 46, 1251–72. Hawtin, M. (1996) Assessing housing needs. In: Needs Assessments in Public Policy (ed. J. Percy-Smith), 98–116. Open University Press, Buckingham. Heal, G.M. (1973) The Theory of Economic Planning. North Holland, Amsterdam. Healey, P., Purdue, M. & Ennis, F. (1993) Gains from Planning? Dealing with the impacts of development. Joseph Rowntree Foundation, York. Hendershott, P.H. (1980) Real user costs and the demand for single family housing. Brookings Papers on Economic Activity, 401–42. Hendershott, P.H. (1988) Home ownership and real house prices: Sources of change, 1965–1985. Housing Finance Review, 7, 1–18. Hendershott, P.H. & Hu, S.C. (1981) Inﬂation and extraordinary returns on owneroccupied housing: Some implications for capital allocation and productivity growth. Journal of Macroeconomics, 3 (2), 177–203. Hendershott, P.H. & Ling, D.C. (1984) Prospective changes in tax law and the value of depreciable real estate. AREUEA Journal, 12 (3), 297–317. Hendershott, P.H. & Shilling, J.D. (1982) Capital allocation and the economic recovery Tax Act of 1981. Public Finance Quarterly, 10 (2), 242–73. Hendershott, P.H. & Slemrod, J. (1983) Taxes and the User Cost of Capital for OwnerOccupied Housing. American Real Estate and Urban Economics Association Journal, 10 (4), 375–93. Hendershott, P.H., Follain, J.R. & Ling, D.C. (1987) Effects on real estate. In: Tax Reform and the US Economy (ed. J.A. Pechman), 71–102. Brookings Institute, Washington, DC. Henderson, J.V. (1985) Economic Theory and the Cities. Academic Press, New York. Henderson, J.V. & Ioannides, Y. (1983) A model of housing tenure choice. American Economic Review, 73, 98–113. Henderson, J.V. & Ioannides, Y. (1987) Owner-occupancy: Investment vs. consumption demand. Journal of Urban Economics, 21, 228–41. Henley, A. (1998) Residential mobility, housing equity and the labour market. Economic Journal, 108, 414–27. Hicks, J.R. (1939) Value and Capital. Clarendon Press, Oxford. Hicks, J.R. (1960) Linear Theory. Economic Journal, 70, 671–709. Hill, R.C., Sirmans, C.F. & Knight, J.R. (1999) A random walk down Main Street? Regional Science and Urban Economics, 29 (1), 89–103. Hills, J. (1991) Unravelling Housing Finance: Subsidies, Beneﬁts and Taxation. Clarendon Press, Oxford.

References

299

Hills, J. (1995) Inquiry into Incomes and Wealth, vol. 2: A summary of the evidence. Joseph Rowntree Foundation, York. Hills, J. (2000) Reinventing Social Housing Finance. IPPR, London. Hills, J., Hubert, F., Tomann, H. & Whitehead, C. (1990) Shifting subsidies from bricks and mortar to people. Housing Studies, 5, 147–67. Hirsch, F. (1976) Social Limits to Growth. Harvard University Press, Cambridge, Mass. Hirshleifer, J. (1970) Investment, Interest and Capital. Prentice Hall, Englewood Cliffs, NJ. Hocking, R.R. (1976) The analysis and selection of variables in linear regression. Biometrics, 32, 1–49. Hoffman, J. & Kurz, C. (2002) Rent Indices for Housing in West Germany: 1985 to 1998. European Central Bank Working Paper No. 116. Holmans, A. (1996) A decline in young owner-occupiers in the 1990s. Housing Finance, 29, 13–20. Holmans, A. (2001) Housing Demand and Need in England 1996–2016. Town and Country Planning Association, London. Hölmstrom, B. (1979) Moral hazard and observability. Bell Journal of Economics, 10, 74–91. Hölmstrom, B. (1982) Moral hazard in teams. Bell Journal of Economics, 13, 324–40. Holt, C.C. & Shelton, J.P. (1962) The lock-in effect of the Capital Gains Tax. National Tax Journal, XV, (4), 337–52. Hopkins, C. Benjamin, C. & Carter, A. (1997) Regeneration: Some Legal and Practical Issues. Lawrence Graham, BURA. Hort, K. (1997) On price formation and quantity adjustment in Swedish housing markets. Economic Studies Number 34. University of Uppsala, Sweden. Housing Research Findings no. 19 (1990) Survey of 10,000 households in six conurbations. Nov. 1990, Joseph Rowntree Foundation, York. Housing Studies (2000) Special issue: Housing and Health. Housing Studies, 15, 3. Housing Studies (2002) Special Issue: Opportunity, Deprivation and the Housing Nexus: Trans-Atlantic Perspectives. Housing Studies, 17, 1. Hoyt, Homer (1939) The Structure and Growth of Residential Neighborhoods in American Cities. Federal Housing Administration, Washington, D.C. Hsiao, C. (1986) Analysis of Panel Data. Econometric Society Monographs. Cambridge University Press, Cambridge. Hughes, G.A. & McCormick, B. (1981) Do council house policies reduce migration between regions? Economic Journal, 91, 919–39. Hughes, G.A. & McCormick, B. (1985) Migration intentions in the UK. Which households want to migrate and which succeed? Economic Journal, 95 (Suppl.), 113–23. Hughes, G.A. & McCormick, B. (1987) Housing markets, unemployment and labour market ﬂexibility in the U.K. European Economic Review, 31, 615–45. Hughes, G.A. & McCormick, B. (1990) Housing and labour market mobility. In: Housing and the National Economy (ed. J. Ermisch). Avebury, Aldershot.

300

References

Hughes, G.A. & McCormick, B. (2000) Housing Policy and Labour Market Performance. DETR, London. Hunter, A. (1974) Symbolic Communities. University of Chicago Press, Chicago, IL. Hunter, A. (1979) The urban neighborhood: its analytical and social contexts. Urban Affairs Quarterly, 14 (3), 267–88. Industrial Systems Research (1999) Political barriers to housebuilding in Britain: a critical case study of protectionism and its industrial-commercial effects. ISR Business and Political-Legal Environment Reports, ISR, Manchester. Ingram, G. (1979) Simulation and econometric approaches to modelling urban areas. In: Current Issues in Urban Economics (eds M. Straszheim & P. Mieszkowski). Johns Hopkins University Press, Baltimore. Jaffe, A.J. (1996) On the role of transaction costs and property rights in housing markets. Housing Studies, 11 (3), 425–35. Jones, C. (2002) The deﬁnition of housing market areas and strategic planning. Urban Studies, 34, (Forthcoming). Jones, C. & Murie, A. (1999) Reviewing the Right to Buy. University of Birmingham Press. Jones, C. & Watkins, C. (1999) Planning and the Housing System. In: Planning Beyond 2000 (ed. P. Allmendinger & M. Chapman), 89–104. Wiley, Chichester. Jones, C., Leishman, C. & Watkins, C. (2000) Structural change in local urban housing markets. Paper presented at the American Real Estate Society Conference, Santa Barbara, California, March. Jones, C., Leishman, C. & Watkins, C. (2001a) Migration Linkages Between Urban Housing Submarkets: Theory and Evidence. (Unpublished). Jones, C., Leishman, C. & Watkins, C. (2001b) Housing market processes, urban housing sub markets and planning policy. Paper presented at Royal Institute of Chartered Surveyors ‘Cutting Edge’ Conference, Oxford, U.K., September. Jones, L.D. (1989) Current wealth and tenure choice. AREUEA Journal, 17 (1), 7–40. Jones, L.D. (1995) Testing the central prediction of housing tenure transition models. Journal of Urban Economics, 38, 50–73. Joseph Rowntree Foundation (1994) Inquiry into planning for housing. Report. Joseph Rowntree Foundation, York. Jud, G.D. & Seaks, T.G. (1994) Sample selection bias in estimating housing sales prices. Journal of Real Estate Research, 9(3), 289–98. Kain, J. & Agpar, W. (1985) Housing and Neighborhood Dynamics. Harvard University Press, Cambridge, Massachusetts. Kain, J.F. & Quigley, J. (1972a) Housing market discrimination, homeownership, and savings behavior. American Economic Review, 62 (3), 263–77. Kain, J.F. & Quigley, J.M. (1972b) Note on owners’ estimates of housing value. Journal of the American Statistical Association, 67, 803–6. Kaufman, D. & Quigley, J. (1987) The consumption beneﬁts of investment in infrastructure: The evaluation of sites and services programs in underdeveloped countries. Journal of Development Economics, 25, 263–84.

References

301

Keeble, D. & Walker, S. (1994) New ﬁrms, small ﬁrms and dead ﬁrms: Spatial patterns and determinants in the United Kingdom. Regional Studies, 28, 411–28. Kellet, J. (1989) Health and housing. Journal of Psychosomatic Research, 33, 255–68. Kennedy, P.E. (1981) Estimation with correctly interpreted dummy variables in semi logarithmic equations. American Economic Review, 71, 801. Kerr, M. (1988) The Right to Buy: a National Survey of Tenants and Buyers of Council Houses. HMSO, London. Kiefer, D.M. (1978) The equity of alternative policies for the Australian homeowner. The Economic Record, 54, 127–39. Kiefer, D.M. (1980) The interaction of inﬂation and the U.S. income tax subsidies of housing. National Tax Journal, 34 (4), 433–45. King, A.T. (1975) The demand for housing: Integrating the roles of journey-to-work, neighborhood quality, and prices. In: Household Production and Consumption (ed. N. Terlecky). NBER. King, A.T. (1977) Estimating property tax capitalization: A critical comment. Journal of Political Economy, 85 (2), 425–31. King, M.A. (1981) An econometric model of tenure choice and demand for housing as a joint decision. Journal of Public Economics, 14, 137–59. King, M. (1990) Discussion of J. Muellbauer and A. Murphy; is the U.K. balance of payments sustainable? Economic Policy, 11, 383–87. Kleinman, M.P. & Whitehead, C.M.E. (1991) Setting a Rent Structure: A Handbook for Social Landlords. Scottish Homes, Glasgow. Knight, J.R., Dombrow J. & Sirmans, C.F. (1995) A varying parameters approach to constructing house price indexes. Real Estate Economics, 23 (2), 187–205. Knight, J.R., Carter Hill, R. & Sirmans, C.F. (1992) Biased prediction of housing values. AREUEA Journal, 20 (3), 427–56. Koskela, E. & Viren, M. (1992) Inﬂation, capital markets and household saving in the Nordic countries. Scandinavian Journal of Economics, 94, 215–27. Kotlikoff, L. (1988) Intergenerational transfers and savings. Journal of Economic Perspectives, 2, 41–58. Laidler, D. (1969) Income tax incentives for owner-occupied housing. In: Taxation of Income from Capital (eds A.C. Harberger & M. J. Bailey), 50–76. Brookings Institute, Washington, D.C. Lambert, C. & Bramley, G. (2002) Managing urban development: Land use planning and city competitiveness. In: Urban Competitiveness: Policies for dynamic cities (ed. I. Begg). Policy Press, Bristol. Lancaster, K. (1966) A new approach to consumer theory. Journal of Political Economy, 74, 132–57. Lancaster, K. (1971) Consumer Demand: A New Approach. Columbia University Press, New York. Leamer, E.E. (1978) Speciﬁcation Searches: Ad Hoc Inference With Nonexperimental Data. Wiley, New York. Leishman, C. & Bramley, G. (forthcoming) A Local Housing Market Model with Spatial Interaction and Land Use Planning Controls. Under review, .

302

References

Leven, C., Little, J., Nourse, H., & Reed, R. (1976) Neighborhood Change: The Dynamics of Urban Decay. Praeger, New York. Levine, D.N. (Ed.) (1971) George Simmel: On Individuality and Social Forms. University of Chicago Press, Chicago. Ling, D.C. (1992) Real estate values, Federal income taxation and the importance of local market conditions. Journal of the American Real Estate and Urban Economic Association, 20 (1), 122–39. Ling, D.C. & McGill, G.A. (1992) Measuring the size and distributional effects on homeowner tax preferences. Journal of Housing Research, 3 (2), 273–304. Linneman, P. (1981) The demand for residence site characteristics. Journal of Urban Economics, 9, 129–48. Linneman, P. (1985) An economic analysis of the homeownership decision. Journal of Urban Economics, 17 (2), 230–46. Linneman, P. & Megbolugbe, I.F. (1992) Housing affordability: Myth or reality? Urban Studies, 29 (1), 369–92. Linneman, P. & Wachter, S. (1989) The impacts of borrowing constraints on homeownership. AREUEA Journal, 17 (4), 389–402. Litzenberger, R.H. & Sosin, H.B. (1978) Taxation and the incidence of homeownership across income groups. Journal of Finance, XXXIII (3), 947–61. Liu, C.H., Grisson, T.V. & Hartzell, D.J. (1995) Superior real estate investment performance: Enigma or illusion? A critical review of the literature. In: Alternative ideas in real estate investment (eds A.L. Schwartz, Jr. & D.D. Kaplin). Kluwer, Boston. Llewellyn-Davies (1994) Providing more homes in urban areas. Report to Joseph Rowntree Foundation. SAUS Publications (now the Policy Press), Bristol. Lowry, I. (1960) Filtering and housing standards: A conceptual analysis. Land Economics, 36, 362–70. Lusht, K.M. (1976) The use of design stratiﬁcation in mass appraisal. Assessors Journal, March, 47–57. Lynn, P. (1991) The Right to Buy: a National Follow-up Survey of Tenants of Council Homes in England. HMSO, London. Macho-Stadler, I. & Pérez-Castrillo, D. (1997) An Introduction to the Economics of Information. Oxford University Press, Oxford. Maclennan, D. (1977a) The economics of selling council houses in the Scottish housing system. Occasional Paper 77–104, Department of Political Economy, University of Aberdeen. Maclennan, D. (1977b) Some thoughts on the nature and purpose of house price studies. Urban Studies, 14, 39–71. Maclennan, D. (1982) Housing Economics: An Applied Approach. Longmans, London and New York. Maclennan, D. (1989) Housing in Scotland, 1977–87. In: Guide to Housing, 3rd edn. (ed. M. E. H. Smith), 671–704. The Housing Centre Trust, London. Maclennan, D. (1986a) The Demand for Housing: economic perspectives and planning practices. Scottish Development Department, Edinburgh. Maclennan, D. (1986b) The rents of public housing in Britain. In: Papers for the Duke of Edinburgh’s Inquiry into British Housing (ed. R. Best). NFHA, London.

References

303

Maclennan, D. (1986c) The pricing of public housing in the United Kingdom. Inquiry into British Housing: Supplement. National Federation of Housing Associations, London. Maclennan, D. (1991) Extending the strategic role. In: The Housing Service of the Future (eds D. Donnison & D. Maclennan), 185–211. Institute of Housing and Longman, Coventry and Harlow. Maclennan, D. (1993a) Housing policy and economic recovery: housing and the economy – what next? Housing policy and economic recovery, Brieﬁngs 2. Joseph Rowntree Foundation, York. Maclennan, D. (1993b) Spillovers, Expectations and Residents Beneﬁts in a Housing Revitalisation Programme: Glasgow 1977–1987. Tidjschrift voor Economische en Sociale Geograﬁe, 84, 294–303. Maclennan, D. (1994) A competitive U.K. economy: the challenges for housing policy. Housing Research Summary, Joseph Rowntree Foundation, York. Maclennan, D. (2000) Changing places, engaging people. Joseph Rowntree Foundation, York. Maclennan, D. (2002) Recognising place and managing territory. (Unpublished paper). Maclennan, D. & Gibb, K. (1993) Housing Indicators and Research for Policy from the Perspective of Applied Economics. Netherlands Journal of Housing and the Built Environment, 8(1), 49–60. Maclennan, D. & Gibb, K. (1994) Modelling metropolitan housing systems and seeking new micro foundations: big ideas in a cold climate? Paper presented to the seminar ‘Housing Policy and Research Priorities’. Centre for Housing Research, Glasgow University, Glasgow, 22–24 March. Maclennan, D. & Meen, G. (1993) Housing markets and national economic performance in OECD countries: lessons for the UK. Housing Policy and Economic Recovery, Brieﬁngs 3. Joseph Rowntree Foundation, York. Maclennan, D. & O’Sullivan, A.J. (1987) Housing Policy in the United Kingdom: Efﬁcient or Equitable? In: Housing Markets and Policies under Fiscal Austerity (ed. W. van Vliet). Greenwood Press, Westport. Maclennan, D. & Tu, Y. (1996) Economic perspectives on the structure of local housing markets. Housing Studies, 11, 387–406. Maclennan, D. & Williams, R. (1990a) Affordable Housing in Europe. Joseph Rowntree Foundation, York. Maclennan, D. & Williams, R. (1990b) Housing Subsidies and the Market: An International Perspective. Joseph Rowntree Foundation, York. Maclennan, D. & Wood, G.A. (1982a) Information acquisition: Patterns and strategies. In: Modelling Housing Market Search (ed. W. A. V. Clark), 134–59. Croom Helm, London. Maclennan D. & Wood, G.A. (1982b) The Pricing of Public Housing: Principles and Implementation. Northern Ireland Housing Executive. Maclennan, D., Munro, M. & Wood, G.A. (1987) Housing choice and the structure of urban housing markets. In: Between State and Market Housing in the Postindustrial Era (eds B. Turner, J. Kemeny & L. Lundquist), 26–52. Almquist and Hicksell International, Gothenburg.

304

References

Maclennan, D., Gibb, K., & More, E.A. (1990) Paying for Britain’s Housing. Joseph Rowntree Foundation, York. Maclennan, D., Gibb, K. & More, E.A. (1991a) Fairer Subsidies, Faster Growth. Joseph Rowntree Foundation, York. Maclennan, D., More, E.A. & Gibb, K.D. (1991b) Subsidising home-ownership. Housing Research Findings No 35, May 1991. Joseph Rowntree Foundation, York. Maclennan, D., More, E.A. & Gibb, K.D. (1991c) Subsidising the independent rented sector. Housing Research Findings No 37, May 1991. Joseph Rowntree Foundation, York. Maclennan, D., More, E.A. & Gibb, K.D. (1991d) Inconsistent rent-setting for council housing. Housing Research Findings No 36, May 1991. Joseph Rowntree Foundation, York. Maclennan, D., More, A. & Munro, M. (1994) Analysing local housing systems: new wine, new bottles or new tastes? Paper presented to the seminar ‘Housing Policy and Research Priorities’. Centre for Housing Research and Urban Studies, Glasgow University, Glasgow, 22–24 March. Maclennan, D., Meen, G., Stephens, M. & Gibb, K. (1997) Fixed Commitments, Uncertain Incomes: Sustainable Owner Occupation and the Economy. Joseph Rowntree Foundation, York. Maclennan, D., More, A., O’Sullivan, A. & Young, G. (1998a) Local Housing Systems Analysis: best practice guide. Scottish Homes, Edinburgh. Maclennan, D., Muellbauer, J. & Stephens, M. (1998b) Asymmetries in housing and ﬁnancial market institutions and EMU. Oxford Review of Economic Policy, 14, 54–80. Maclennan, D., O’Sullivan, A. & Macintyre, C. (2000) Evolving the Right to Buy: Evidence for Scotland. Scottish Executive, Edinburgh. MacNee, K. (1993) The Right to Buy in Scotland: an Assessment of the Impact of the First Decade of the Right to Buy. Scottish Ofﬁce Central Research Unit, Edinburgh. MacNevin, A.S. (1997) Tax Effects on Rental Housing in Halifax, Nova Scotia. Canadian Tax Journal, 45, (1), 87–113. Maddala, G.S. (1983) Limited Dependent and Qualitative Variables in Econometrics. Cambridge University Press. Malatesta, P.H. & Hess, E. (1986) Discount mortgage ﬁnancing and housing prices. Housing Finance Review, 5, 25–41. Malpass, P. (1980) Council house sales in Yeovil District. Policy and Politics, 8, 308–15. Malpass, P. (1983) Residualization and the restructuring of housing tenure. Housing Review, 32, 44–45. Malpezzi, S. (1990) Urban housing and ﬁnancial markets: some international comparisons. Urban Studies, 27 (6), 971–1022. Malpezzi, S. (1996) Housing prices, externalities and regulation in U.S. metropolitan areas. Journal of Housing Research, 7 (2), 209–41.

References

305

Malpezzi, S. (1998) Welfare analysis of rent control with side payments: A natural experiment in Cairo, Egypt. Regional Science and Urban Economics, 28 (6), 773–96. Malpezzi, S. (1999) A simple error correction model of house prices. Journal of Housing Economics, 8, 27–62. Malpezzi, S. (2000) Housing. In: Designing Household Survey Questionnaires for Developing Countries: Lessons from Fifteen Years of the Living Standards Measurement Study (eds M. Grosh & P. Glewwe). Oxford University Press. Malpezzi, S. & Green, R.K. (1996) What has happened to the bottom of the U.S. housing market? Urban Studies, 33 (10), 1807–20. Malpezzi, S. & Mayo, S.K. (1987) The demand for housing in developing countries. Economic Development and Cultural Change, 35 (4), 687–721. Malpezzi, S. & Mayo, S.K. (1994) A Model Design for a Developing Country Housing Market Study. University of Wisconsin-Madison, Center for Urban Land Economics Research Working Paper. Malpezzi, S. & Mayo, S.K. (1997) Housing and urban development indicators: A good idea whose time has returned. Real Estate Economics, 25 (1), 1–11. Malpezzi, S., Ozanne, L. & Thibodeau, T. (1980) Characteristic Prices of Housing in 59 SMSAs. The Urban Institute, Washington, DC. Malpezzi, S., Ozanne, L. & Thibodeau, T. (1987) Microeconomic estimates of housing depreciation. Land Economics, 63(4), 373–85. Malpezzi, S., Chun, G. & Green, R. (1998) New place-to-place housing price indexes for U.S. metropolitan areas, and their determinants: An application of housing indicators. Real Estate Economics, 26 (2), 235–75. Mankiw, N.G. & Weil, D.N. (1989) Baby boom, baby bust and the housing market. Regional Science and Urban Economics, 19, 235–58. Marmot, M. & Wilkinson, R. (eds) (1999) Social Determinants of Health. University Press, Oxford. Marris, P. (1997) Witnesses, Engineers, and Story Tellers. University of Maryland, Urban Studies and Planning Program. Mason, C. & Quigley, J.M. (1996) Non-parametric housing prices. Housing Studies, 11 (3), 373–85. Mayer, C. & Somerville, T. (2000) Land use regulation and new construction. Regional Science and Urban Economics, 30, 639–62. Mayo, S.K. (1981) Theory and estimation in the economics of housing demand. Journal of Urban Economics, 10, 95–116. Mayo, S.K. & Gross, D.J. (1987) Sites and services and subsidies: The economics of low cost housing in developing countries. World Bank Economic Review, 1 (2), 301–35. McBeath, J. (1997) Learning, school and neighbourhood. Paper presented at seminar on ‘The Wider Impacts of Housing’, Scottish Homes. McCarthy, G., Van Zandt, S. & Rohe, W. (2001) The economic beneﬁts and costs of homeownership; a critical assessment of the research. Research Institute for Housing America, Working Paper No. 01–02. McClure, K. (2000) The low-income housing tax credit as an aid to housing ﬁnance: How well has it worked? Housing Policy Debate, 11 (1), 91–114.

306

References

McDonald, J. (1997) Fundamentals of Urban Economics. Prentice-Hall, Upper Saddle River, NJ. McFadden, D. (1978) Modelling the choice of residential location. In: Planning Models (ed. A. Karlquist), 75–96. North Holland, Amsterdam. Meen, G.P. (1990) The removal of mortgage market constraints and the implications for econometric modelling of U.K. house prices. Oxford Bulletin of Economics and Statistics, 52, 1–24. Meen, G. (1993) The treatment of house prices in macroeconomic models: a comparison exercise. DOE Occasional Paper, Housing and Urban Monitoring Analysis. Meen, G.P. (1996) Ten propositions in U.K. housing macroeconomics: An overview of the eighties and early nineties. Urban Studies, 33, 425–44. Meen, G.P. (1998) Modelling sustainable home-ownership: Demographics or economics? Urban Studies, 35, 1919–34. Meen, G.P. (1999a) Models of housing in London and the South East. Paper presented at Citywide seminar, April, Glasgow Clyde Port Authority. Meen, G.P. (1999b) Regional house prices and the ripple effect: A new interpretation. Housing Studies, 14, 733–53. Meen, G.P. (2000) Housing cycles and efﬁciency. Scottish Journal of Political Economy, 47, 114–40. Meen, G.P. (2001) Modelling spatial housing markets: Theory, analysis and policy. Kluwer Academic Publishers, Boston. Meen, G.P. & Andrew, M. (1998) On the aggregate housing market implications of labour market change. Scottish Journal of Political Economy, 45, 393–419. Meen, G.P. & Andrew, M. (1999) Spatial structure and social exclusion. Discussion Papers in Urban and Regional Economics 140: Department of Economics, University of Reading. Meen, G.P., Gibb, K., Mackay, D. & White, M. (2001) The Economic Role of New Housing. NHBC, London. Meese, R. & Wallace, N. (1991) Nonparametric estimation of dynamic hedonic price models and the construction of residential house price indices. Journal of American Real Estate and Urban Economics Association, 19, 308–32. Meese, R.A. & Wallace, N.E. (1994) Testing the present value relation for housing prices: Should I leave my house in San Francisco? Journal of Urban Economics, 35, 245–66. Memery, C. (2001) The housing system and the Celtic Tiger: the state response to a housing crisis of affordability and access. European Journal of Housing Policy, 1 (1), 79–104. Michaels, R. & Smith, V.K. (1990) Market segmentation and valuing amenities with hedonic models: the case of hazardous waste sites. Journal of Urban Economics, 28, 223–42. Miles, D. (1997) A household level study of the determinants of incomes and consumption. Economic Journal, 107, 1–25. Mills, E. (1967) An aggregative model of resource allocation in a metropolitan area. American Economic Review, 57, 197–210.

References

307

Mills, E. (1972) Studies in the Structure of the Urban Economy. Johns Hopkins University Press, Baltimore. Mills, E.S. (1987) Has the United States over-invested in housing? Journal of the American Real Estate and Urban Economics Association, 15 (1), 601–16. Mills, E.S. & Simenauer, R. (1996) New hedonic estimates of regional constant quality housing prices. Journal of Urban Economics, 39 (2), 209–15. Minford, P., Ashton, P. & Peel, M. (1988) The effects of housing distortions on unemployment. Oxford Economic Papers, 40, 322–45. Ministry of Reconstruction (1945) Housing. Cd 6609, HMSO, London. Monk, S. (2000a) The ‘key worker’ problem: The link between employment and housing. In: Restructuring Housing Systems (eds S. Monk & C. M. E. Whitehead). Joseph Rowntree Foundation, York. Monk, S. (2000b) The use of price in planning for housing. In: Restructuring Housing Systems (eds S. Monk & C. M.E. Whitehead), 198–207. Joseph Rowntree Foundation, York. Monk, S. & Whitehead, C.M.E. (2000) The use of housing and land prices as a planning tool: a summary document. Cambridge Housing and Planning Research, Research Report 1, Cambridge. Monk, S., Pearce, B. & Whitehead, C. (1991) Planning, land supply and house prices: a literature review. Monograph 21. Department of Land Economy, University of Cambridge. Granta Publications, London & Cambridge. Monk, S., Pearce, B., & Whitehead, C.M.E. (1996) Land-use planning, land supply, and house prices. Environment & Planning A, 28, 495–511. More, A. (2002) Planning and Strategy. In: Introduction to Management, 2nd edn. (ed. D. Boddy). (Forthcoming, Pearson). Moulton, B.R. (1995) Inter-area indexes of the cost of shelter using hedonic quality adjustment techniques. Journal of Econometrics, 68 (1), 181–204. Mozolin, M. (1994) The geography of housing values in the transformation to a market economy – A case study of Moscow. Urban Geography, 15 (2), 107–27. Muellbauer, J. (1990) The housing market and the U.K. economy: Problems and opportunities. In: Housing and the National Economy (ed. J. Ermisch). Avebury, Aldershot. Muellbauer, J. & Murphy, A. (1990) Is the U.K. balance of payments sustainable? Economic Policy, 11, 347–82. Muellbauer, J. & Murphy, A. (1997) Booms and busts in the U.K. housing market. Economic Journal, 107, 1701–27. Munnell, A.H., Tootell, G.M.B., Browne, L.E., & McEneaney, J. (1996) Mortgage Lending in Boston: Interpreting HMDA Data. American Economic Review, 86 (1), 25–53. Munro, M. & Tu, Y. (1997) U.K. House Price Dynamics: past and future trends. CML Research Report, Council of Mortgage Lenders, London. Murie, A. (1975) The sale of council houses. Occasional Paper 35, Centre for Urban and Regional Studies, University of Birmingham. Murie, A. (1989) Lost opportunities? Council house sales and housing policy in Britain 1979–89. Working Paper 80, School for Advanced Urban Studies, University of Bristol.

308

References

Murie, A. & Nevin, B. (1997) Beyond a Half Way Housing Policy. NHF, London. Murray, C. (1984) Losing Ground: America’s Social Policy, 1950–1980. Basic Books, New York. Musgrave, R.A. (1959) The Theory of Public Finance. McGraw-Hill, New York. Muth, R. (1969) Cities and Housing. University of Chicago Press, Chicago. Muth, R. (1985) Models of land-use, housing and rent: An evaluation. Journal of Regional Science, 25, 593–606. Narwold, A. (1992) The distribution of the beneﬁts of tax arbitrage in the housing market. Journal of Urban Economics, 32, 367–76. National Council of Social Service (1980) Rural Housing in East Hampshire. NCSS. Needham, B. & Lie, R. (1994) The public regulation of property supply and its effects on private prices, risks and returns. Journal of Property Research, 11, 199–213. Nelson, J.P. (1982a) Estimating demand for product characteristics: Comment. Journal of Consumer Research, 9 (2), 219–20. Nelson, J.P. (1982b) Highway noise and property values: A survey of recent evidence. Journal of Transport Economics and Policy, 16, 117–38. Neutze, M. (1987) The supply of land for a particular use. Urban Studies, 24, 379–88. Nevin, B., Lee, P., Murie, A., Goodson, L. & Phillimore, J. (2001) The West Midlands Housing Markets: Changing Demand, Decentralisation and Urban Regeneration. Centre for Urban and Regional Studies, University of Birmingham. Niskanen, W.A. (1994) Bureaucracy and Public Economics. Edward Elgar, Aldershot. Nordvik, V. (2000) Tenure ﬂexibility and the supply of private rental housing. Regional Science and Urban Economics, 30, 59–76. Nourse, H.O. (1963) The effect of public housing on property values in St Louis. Land Economics, 39, 434–41. Nozick, R. (1974) Anarchy, State, and Utopia. Basil Blackwell, Oxford. Oates, W. (1981) On local ﬁnance and the Tiebout Model. American Economic Review, 71 (2), 93–8. Ohta, M. & Griliches, Z. (1975) Automobile prices revisited: Extensions of the hedonic price hypothesis. In: Household Production and Consumption, Studies in Income and Wealth. Vol. 40 (ed. N. E. Terleckyj). [325–398.] University of Chicago Press, for the National Bureau of Economic Research, Chicago. Olsen, E.O. (1968) A competitive theory of the housing market. American Economic Review, 58, 612–22. Olsen, E.O. (1972) An econometric analysis of rent control. Journal of Political Economy, 80, 1081–1100. Olsen, E.O. (1987) The demand and supply of housing services: A critical review of the empirical literature. In: Handbook of Regional and Urban Economics (ed. E.S. Mills), Vol. 2, Elsevier. Olsen, E.O. & Barton, D.M. (1983) The beneﬁts and costs of public housing in New York City. Journal of Public Economics, 20, 299–332. O’Sullivan, A. (1984) Misconceptions in the current housing subsidy debate. Policy and Politics, 12, 119–44.

References

309

O’Sullivan, A. (1987) The Deﬁnition and Measurement of Housing Subsidies for the U.K., 1977. Unpublished doctoral thesis, University of Sussex. Oswald, A.J. (1996) A conjecture on the explanation for high unemployment in the industrialised nations. University of Warwick Working Paper. Oswald, A.J. (1997a) Thoughts on NAIRU. Correspondence to Journal of Economic Perspectives, 11, 227–28. Oswald, A.J. (1997b) The missing piece of the unemployment puzzle. An inaugural lecture. Unpublished. Oswald, A.J. (1999) The housing market and Europe’s unemployment: a non-technical paper. University of Warwick Working Paper. Ozanne, L. & Malpezzi, S. (1985) The efﬁcacy of hedonic estimation with the annual housing survey: Evidence from the demand experiment. Journal of Economic and Social Measurement, 13 (2), 153–72. Pace, R.K. (1993) Nonparametric methods with applications to hedonic models. Journal of Real Estate Finance and Economics, 7 (3), 185–204. Pace, R.K. & Gilley, O.W. (1990) Estimation employing a priori information within mass appraisal and hedonic pricing models. Journal of Real Estate Finance and Economics, 3 (1), 55–72. Pace, R.K. & Gilley, O.W. (1997) Using the spatial conﬁguration of the data to improve estimation. Journal of Real Estate Finance and Economics, 14 (3), 333–40. Palm, R. (1978) Spatial segmentation of the urban housing market. Economic Geography, 54, 210–21. Paris, C. & Blackaby, B. (1979) Not Much Improvement, Heinemann, London. Pasha, H.A. & Butts, M.S. (1996) Demand for housing attributes in developing countries: A case study of Pakistan. Urban Studies, 33 (7), 1141–54. Pawson, H. & Watkins, C. (1999) Resale of former public sector homes in rural Scotland. Scottish Geographical Magazine, 114, 157–63. Pawson, H., Watkins, C. & Morgan, J. (1997) Right to Buy Resales in Scotland. Scottish Ofﬁce, Edinburgh. Pesaran, M.H. & Smith, R. (1995) Estimating long-run relationships from dynamic heterogeneous panels. Journal of Econometrics, 68, 79–113. Peterson, A.W.A., Pratten, C.F. & Tatch, J. (1998) An Economic Model of Demand and Need for Social Housing. DETR, London. Phillips, R.S. (1988) Unraveling the rent-value puzzle: An empirical investigation. Urban Studies, 25, 487–96. Plaut, S.E. (1987) The timing of housing tenure transition. Journal of Urban Economics, 21, 312–22. Podogzinski, H.U. & Sass, T.R. (1991) Measuring the effects of municipal zoning regulations: a survey. Urban Studies, 28, 597–621. Pollak, R. (1985) A transaction cost approach to families and households. Journal of Economic Literature, 23, 581–608. Posner, R.A. (1972) The appropriate scope of regulation in the cable television industry. The Bell Journal of Economics and Management Science, 3, 98–129. Potepan, M.J. (1989) Interest rates, income, and home improvement decisions. Journal of Urban Economics, 25, 282–94.

310

References

Poterba, J. (1991) House price dynamics: The role of tax policy and demography. Brookings Papers on Economic Activity, 143–99. Priemus, H. (1992) Housing indicators: An instrument in international housing policy? Netherlands Journal of Housing and the Built Environment, 7 (3), 217–38. Pryce, G. (1999) Construction elasticities and land availability: a two-stage least squares model of housing supply using the variable elasticity approach. Urban Studies, 36 (13), 2283–2304. Quercia, R.G. & Galster, G.C. (2000) Threshold effects and neighborhood change. Journal of Planning Education and Research, 20, 146–63. Quigley, J.M. (1979) What have we learned about urban housing markets? In: Current Issues in Urban Economics (eds M. Straszheim & P. Mieskowski). The Johns Hopkins University Press, Baltimore. Quigley, J.M. (1982a) Nonlinear budget constraints and consumer demand: An application to public programs for residential housing. Journal of Urban Economics, 12, 177–201. Quigley, J.M. (1982b) Estimates of a more general model for consumer choice in the housing market. In: The Urban Economy and Housing (ed. R. Grieson), 125–40. Heath/Lexington, Lexington, Mass. Quigley, J.M. (1987) Interest rate variations, income, and home improvement decisions. Review of Economics and Statistics, 69 (4), 636–43. Quigley, J.M. (1995) A simple hybrid model for estimating real estate price indexes. Journal of Housing Economics, 4 (1), 1–12. Quigley, J.M. (1996) Mortgage performance and housing market discrimination. Cityscope: A Journal of Policy Development and Research, 2 (1), 59–64. Quigley, J.M. (ed.) (1998) The Economics of Housing Markets. Edward Elgar, Cheltenham. Quigley, J.M. (2002) Homeowner mobility and mortgage interest rates: New evidence from the 1990s. Real Estate Economics. Forthcoming. Ram, R. (1986) Government size and economic growth: A new framework and some evidence from cross-section and time-series data. American Economic Review, 76, 191–203. Richmond, P. (1980) The sale of council houses in Worcester. Policy and Politics, 8, 316–17. Robertson, D.S. (1998) Pulling in opposite directions: the failure of post war planning to regenerate Glasgow. Planning Perspectives, 13, 53–67. Robinson, R. (1979) Housing Economics and Public Policy. Macmillan, Basingstoke. Robinson, R. (1981) Housing tax expenditure, subsidies and the distribution of income. The Manchester School, 49, 91–110. Rodgers, B. & Pryor, J. (1998) Divorce and Separation: the outcomes for children. Joseph Rowntree Foundation, York. Rohe, W., McCarthy, G. & Van Zandt, S. (2000) The social beneﬁts and costs of homeownership; a critical assessment of the research. Research Institute for Housing America, Working Paper No. 01–01.

References

311

Rosen, H.S. (1979) Housing decisions and the U.S. income tax: An econometric analysis. Journal of Public Economics, 11, 1–23. Rosen, H.S. (1985) Housing subsidies: Effects on decisions, efﬁciency, and equity. In: Handbook of Public Economics (eds A.T. Aurbach & M.S. Feldstein), vol. 1, 375–420. Holland, New York. Rosen, H.S. & Rosen, K.T. (1980) Federal taxes and homeownership: Evidence from Time Series. Journal of Political Economy, 88 (1), 59–75. Rosen, H.S., Rosen, K.T. & Holtz-Eakin, D. (1984) Housing tenure, uncertainty, and taxation. Review of Economics and Statistics, 66 (3), 405–16. Rosen, S. (1974) Hedonic prices and implicit markets: product differentiation in pure competition. Journal of Political Economy, 82, 34–55. Rothenberg, J. (1967) Economic Evaluation of Urban Renewal. Brookings Institute, Washington, DC. Rothenberg, J., Galster, G., Butler, R., & Pitkin, J. (1991) The Maze of Urban Housing Markets: Theory, Practice and Evidence. University of Chicago Press, Chicago. Rothschild, N.M.V. (1971) The organisation and management of government R&D. A Framework for Government Research and Development (The Rothschild Report), Cmnd.4814, HMSO, London. Rouwendal, J. (1992) The hedonic price function as an envelope of bid functions: An exercise in applied economic theory. Netherlands Journal of Housing and the Built Environment, 7(1), 59–80. Royal Commission on Environmental Pollution (1994) Transport and the Environment. Eighteenth Report. Cmnd. 2674, HMSO, London. Rydin, Y. (1985) Residential Development and the Planning System. Pergamon, Oxford. Rydin, Y. (1986) Housing Land Policy. Gower, Aldershot. Satsangi, M. (1991) Fair’s Fair? A Measurement of Subsidies to Tenants of Glasgow’s Housing Associations. University of Glasgow, Centre for Housing Research. Unpublished. Schelling, T.C. (1971) Dynamic models of segregation. Journal of Mathematical Sociology, 1, 143–86. Schelling, T.C. (1978) Micromotives and Macrobehavior. Norton, New York. Schnare, A.B. & MacRae, C.D. (1975) A model of neighborhood change. The Urban Institute, Contract Report no. 225–4, Washington, DC. Schnare, A. & Struyk, R.J. (1976) Segmentation in urban housing markets. Journal of Urban Economics, 3, 146–66. Schoenberg, S. (1980) Neighborhoods that Work. Rutgers University Press, New Brunswick, NJ. School of Planning & Housing (Edinburgh College of Art) in association with the Department of Building Engineering and Surveying, Heriot-Watt University (2001) The Role of the Planning System in the Provision of Housing. Scottish Executive, Edinburgh. Schwirian, K.P. (1983) Models of neighborhood change. Annual Review of Sociology, 9, 83–102. Scottish Development Department (1977) Assessing Housing Needs: a manual of guidance. Scottish Housing Handbook no. 1, HMSO, Edinburgh.

312

References

Scottish Federation of Housing Associations (2000) Financial Impact of Extending Right to Buy to Housing Associations. SFHA, Edinburgh. Scottish Homes (1996) Sales of Public Sector Dwellings in Scotland 1980–95. Edinburgh. Scottish Housing Advisory Committee (1972) Planning for housing needs: pointers towards a comprehensive approach. HMSO, Edinburgh. Scottish Ofﬁce (1977) Scottish Housing: a consultative document. Cmnd 6852, HMSO, Edinburgh. Sheppard, S. (1999) Hedonic analysis of housing markets. In: Handbook of Regional and Urban Economics (eds P. C. Chesire & E. S. Mills), Vol.3. Elsevier. Shilling, J.D., Sirmans, C.F. & Dombrow, J.F. (1991) Measuring depreciation in single family rental and owner-occupied housing. Journal of Housing Economics, 1(4), 368–83. Short, J., Fleming, S., & Witt, S. (1986) Housebuilding, Planning and Community Action. Routledge, London. Short, K., Garner, T., Johnson, D. & Doyle, P. (1999) Experimental poverty measures: 1990 to 1997. Census Report, 60–205, June. Simpson, D.G. (1998) Why most strategic planning is a waste of time and what you can do about it – Part II. Long Range Planning, 31, 623–27. Smith, G.R. (1999) Area-based initiatives: the rationale and options for area targeting. LSE Case Paper no. 25, London School of Economics, Centre for the Study of Social Exclusion, London. Smith, L., Rosen, K. & Fallis, G. (1988) Recent developments in economic models of housing markets. Journal of Economic Literature, XXVI, 29–64. Smith, M.E.H. (ed.) (1989) Guide to Housing, 3rd edn. The Housing Centre Trust, London. Social Exclusion Unit (1998) Bringing People Together: A National Strategy for Neighbourhood Renewal. HMSO, London. Social Exclusion Unit (2001) A New Commitment to Neighbourhood Renewal: National Strategy Action Plan. January, Cabinet Ofﬁce, London. Soderberg, B. & Janssen, C. (2001) Estimating distance gradients for apartment properties. Urban Studies, 38 (1), 61–79. Solow, R. (1973) On equilibrium models of urban location. In: Essays in Modern Economics (eds M. Parkin & A. Nobay). Longmans, London. Somerville, C.T. & Holmes, C. (2001) Dynamics of the affordable housing stock: Microdata analysis of ﬁltering. Journal of Housing Research, 12 (1), 115–40. Stein, J. (1993) Prices and trading volume in the housing market: A model with downpayment effects. Quarterly Journal of Economics, 110, 379–406. Stephens, M. (2000) Convergence in European mortgage systems before and after EMU. Journal of Housing and the Built Environment, 15, 29–52. Stephens, M., Burns, N. & Mackey, L. (2002) Social market or safety net? Policy Press, Bristol. Stone, E. (1997) All About Homelessness. Shelter, London. Strassman, P.W. (1991) Housing market interventions and mobility: An international comparison. Urban Studies 28, 759–71.

References

313

Straszheim, M. (1974) Hedonic estimation of housing prices: a further comment. Review of Economics and Statistics, 56(3), 404–6. Straszheim, M. (1975) An Econometric Analysis of the Urban Housing Market. NBER, New York. Straszheim, M. (1987) The theory of urban residential location. In: Handbook of Regional and Urban Economics. Volume II Urban Economics (ed. E. Mills). Elsevier Science, BV. Stretton, H. (1971) Ideas for Australian Cities. Georgian House, Melbourne. Stretton, H. (1974) Housing and Government. Australian Broadcasting Corporation, Boyer Lectures. Stretton, H. (1978) Urban Planning in Rich and Poor Countries. Oxford University Press, Oxford. Struyk, R. (with Malpezzi, S. & Wann, F.) (1980) The Performance Funding System Inﬂation Factor: Analysis of Predictive Ability of Candidate Series and Models. Urban Institute Working Paper 1436–02. Suits, D.B. (1984) Dummy variables: mechanics vs. interpretation. Review of Economics and Statistics, 66, 177–80. Suits, D.B., Mason, A. & Chan, L. (1978) Spline Functions Fitted by Standard Regression Methods. Review of Economics and Statistics, 60, 132–9. Sunley, E.M. (1987) Comment on Gordon, R.H., Hines J.R. (Jr), & Summers, L.H. In: The Effects of Taxation on Capital Accumulation (ed. M. Feldstein), 254–57. Chicago University Press, Chicago. Sweeney, J. (1974) A Commodity Hierarchy Model of the Rental Housing Market. Journal of Urban Economics, 1, 288–323. Taub, R.D., Taylor, G. & Dunham, J. (1984) Paths of Neighborhood Change. University of Chicago Press, Chicago. Temkin, K. & Rohe, W. (1996) Neighborhood change and urban policy. Journal of Planning Education and Research, 15, 159–70. Tewdyr-Jones, M. (1996) British Planning Policy in Transition: planning in the 1990s. UCL Press, London. Thibodeau, T.G. (2002) Marking Single-Family Property Values to Market Using Hedonic House Price Equations. Presidential Address to the American Real Estate and Urban Economics Association, Atlanta, January 5. Thomas, A. (1993) The inﬂuence of wages and house prices on British interregional migration decisions. Applied Economics, 25, 1261–68. Tiebout, C. (1956) A pure theory of local expenditures. Journal of Political Economy, 64 (5), 416–24. Titman, S. D. (1982) The effect of anticipated inﬂation on housing market equilibrium. Journal of Finance, 37, 827–42. Tiwari, P. & Hasegawa, H. (2000) House price dynamics in Mumbai, 1989–1995. Review of Urban and Regional Development Studies, 12(2), 149–63. Tracy, J., Schneider, H., & Chan, S. (1999) Are stocks overtaking real estate in household portfolios? Current Issues in Economics and Finance, 5 (5), 1–6. Triplett, J.E. (1974) Consumer Demand and Characteristics of Consumption Goods. BLS Working Paper No. 22.

314

References

Truesdale, D. (1980) House sales and owner occupation in Stevenage New Town. Policy and Politics, 8, 318–23. Tsoukis, C. & Westaway, P. (1991) A forward-looking model of housing construction in the U.K. Econometric Modelling, 11, 266–78. Tu, Y. (1997) The local housing submarket structure and its properties. Urban Studies, 34 (2), 337–53. Tu, Y. & Goldﬁnch, J. (1996) A two stage housing choice forecasting model. Urban Studies, 33 (3), 517–37. Turner, B. (1997) Housing cooperatives in Sweden: The effects of financial deregulation. Journal of Real Estate Finance and Economics, 15 (2), 193–217. Turner, B. & Whitehead, C.M.E. (1993) Housing Finance in the 1990s. Research Report SB:56, The National Swedish Institute for Building Research, Gavle. Turner, B. & Whitehead, C.M.E. (2002) Reducing housing subsidy: Swedish housing policy in an international context. Urban Studies, 39 (2), 201–17. Turner, B., Whitehead, C.M.E. & Jakobsson, J. (1996) Comparative Housing Finance. Swedish Government Housing Commission, Bosdstadspolitik 2000, Expertrapporter, SOU. Vandell, K.D. (1995) Market factors affecting spatial heterogeneity among urban neighborhoods. Housing Policy Debate, 6 (1), 103–39. Vandell, K.D. (2000) Comment on Steven C. Bourassa and William G. Grigsby’s ‘Income Tax Concessions for Owner-Occupied Housing’. Housing Policy Debate, 11 (3), 561–74. Van Zilj, V. (1993) A Guide to Local Housing Needs Assessments. Institute of Housing, Coventry. Varady, D.P. (1996) Local Housing Plans: Learning from Great Britain. Housing Policy Debate, 7, 253–92. Venti, F. & Wise, D.A. (1984) Moving and housing expenditure: Transaction costs and disequilibrium. Journal of Public Economics, 23, 207–43. Von Boventer, E. (1978) Bandwagon effects and product cycles in urban dynamics. Urban Studies, 15, 261–72. Wallace, H.A. (1926) Comparative farmland values in Iowa. Journal of Land and Public Utility Economics, 2, 385–92. Wang, F.T. & Zorn, P.M. (1997) Estimating house price growth with repeat sales data: What’s the aim of the game? Journal of Housing Economics, 6, 93–118. Warren, D. (1975) Black Neighborhoods: An Assessment of Community Power. University of Michigan Press, Ann Arbor, MI. Warren, R. & Warren, D. (1977) The Neighborhood Organizer’s Handbook. Notre Dame University Press, South Bend, IN. Watkins, C. (1998) The definition and identification of housing submarkets. Discussion Paper 98–10, Centre for Property Research, Department of Land Economy, University of Aberdeen, Scotland. Watts, H. (1964) An Introduction to the Theory of Binary Variables (Or, All About Dummies). University of Wisconsin. Unpublished. Weber, M. (1978) Economy and Society. University of California Press, Berkeley, California.

References

315

Weibull, W.J. (1983) A dynamic model of trade frictions and disequilibrium in the housing market. Scandinavian Journal of Economics, 85, 373–92. Weibull, W.J. (1984) A stock ﬂow approach to general equilibrium with trade frictions. Applied Mathematics and Computation, 14, 251–76. Weicher, J.C. (2000) Comment on Steven C. Bourassa and William G. Grigsby’s ‘Income Tax Concessions for Owner-Occupied Housing’. Housing Policy Debate, 11 (3), 547–59. Weinberg, D.J., Friedman, J. & Mayo, S.K. (1981) Intraurban residential mobility: The role of transactions costs, market imperfections, and household disequilibrium. Journal of Urban Economics, 9 (3), 332–48. Weiss, Y. (1978) Capital gains, discriminatory taxes, and the choice between renting and owning a house. Journal of Public Economics, 10 (1), 45–55. Wheaton, W. (1974) A Comparative Static Analysis of Urban Spatial Structure. Journal of Economic Theory, 9, 223–37. Wheaton, W. & DiPasquale, D. (1996) Urban Economics and Real Estate Markets. Prentice Hall, Englewood Cliffs, NJ. Whitehead, C.M.E. (1974) The U.K. Housing Market: An Econometric Model. Gower, Aldershot. Whitehead, C.M.E. (1983) The rationale of government interventions. In: Urban Land Policies: Issues and Opportunities (ed. H. Dunkerley). Oxford University Press, Oxford. Whitehead, C.M.E. (1984) Privatisation and housing. In: Privatisation and the Welfare State (eds J. Le Grand & R. Robinson). Allen & Unwin, London. Whitehead, C.M.E. (1991) From need to affordability: an analysis of U.K. housing objectives. Housing Studies, 28 (6), 871–87. Whitehead, C.M.E. (1993) Privatising housing: An assessment of U.K. experience. Housing Policy Debate, 4 (1), 104–39. Whitehead, C.M.E. (1998) The Beneﬁts of Better Homes. Shelter, London. Whitehead, C.M.E. (1999) Urban Housing Markets: Theory and Policy. In: Handbook of Regional and Urban Economics (eds P. Chesire & E.S. Mills), Vol. 3. Elsevier. Whitehead, C. & Odling-Smee, J. (1975) Long-run equilibrium in urban housing. Urban Studies, 12, 315–18. Wilcox, S. (1998) The numbers game. Inside Housing, February, 17–18. Wilkinson, R. (1996) Unhealthy Societies: Afﬂictions of Inequality. Routledge, London. Williams, N.J. (1993) Homeownership and the Sale of Public Sector Housing in Great Britain. In: Ownership, Control and the Future of Housing Policy (ed. R. Allen Hays), Ch.5. Greenwood Press, Westport. Williams, N.J. & Sewel, J.B. (1987) Council house sales in the rural environment. In: Rural Housing in Scotland (eds B.D. MacGregor, D.S. Robertson & M. Shucksmith), Chap. 7. Aberdeen University Press, Aberdeen. Williams, N.J. & Twine, F.E. (1992) Increasing access or widening choice: the role of resold public sector dwellings in the housing market. Environment and Planning A, 24, 1585–98. Williams, N.J. & Twine, F.E. (1993) Moving on and trading up: the experience of households who resell public sector dwellings. Housing Studies, 8, 60–69.

316

References

Williams, N.J., Sewel, J.B. & Twine, F.E. (1984) The sale of council houses – some empirical evidence. Urban Studies, 21, 439–50. Williams, N.J., Sewel, J.B. & Twine, F.E. (1986) Council house sales and residualisation. Journal of Social Policy, 15, 273–92. Williams, N.J., Sewel, J.B. & Twine, F.E. (1987) Council house sales and the electorate: voting behaviour and ideological implications. Housing Studies, 2, 274–82. Williams, N.J., Sewel, J.B. & Twine, F.E. (1988) Council house sales: an analysis of the factors associated with purchase and implications for the future of public sector housing. Tijdschrift voor Economische en Sociale Geograﬁe, 79, 39–49. Williams, N.J. & Twine, F.E. (1994) Locals, incomers and second homes: the role of resold public sector dwellings in rural Scotland. Scandinavian Housing and Planning Research, 11, 193–209. Williamson, O. (1975) Markets and Hierarchies: Analysis and antitrust implications: A study in the economics of internal organisation. The Free Press, New York. Williamson, O.E. (1979) Transaction costs economics: the governance of contractual relations. Journal of Law and Economics, 22, 233–61. Williamson, O. (1985) The Economic Institutions of Capitalism: Firms, markets and relational contracting. The Free Press, New York. Williamson, O. (1986) Economic Organisation. Wheatsheaf Books, Brighton. Willis, K.G. & Cameron, S.J. (1993) Costs and beneﬁts of housing subsidies in the Newcastle area: A comparison of alternative subsidy deﬁnitions across tenure sectors and income deﬁnitions. In: Housing Finance and Subsidies in Britain, (eds D. Maclennan, & K. Gibb). Avebury, Aldershot. Willis, K.G. & Nicholson, M. (1991) Costs and beneﬁts of housing subsidies to tenants from voluntary and involuntary rent control: A comparison between tenures and income groups. Applied Economics, 23 (6), 1103–15. Wilson, W. J. (1987) The Truly Disadvantaged. University of Chicago Press, Chicago IL. Wilson, W. (1999) The Right to Buy. House of Commons Research Paper 99/36. Wiltshaw, D.G. (1985) The supply of land. Urban Studies, 22, 49–56. Witte, A.D., Sumka, H. & Erekson, J. (1979) An estimate of a structural hedonic price model of the housing market: An application of Rosen’s Theory of Implicit Markets. Econometrica, 47, 1151–72. Wood, G.A. (1988a) The role of housing in economic growth. Working Paper No.21, Murdoch University. Wood, G.A. (1988b) Housing tax expenditures in OECD countries: Economic impacts and prospects for reform. Policy and Politics, 16 (4), 235–50. Wood, G.A. (1992) How do Australian State and Local Governments tax residential housing? Australian Tax Forum, 9 (4), 441–72. Wood, G.A. (1995) The taxation of owner occupied housing in Australia: affordability and distributional issues. PhD dissertation, Murdoch University. Wood, G.A. (1996) The contribution of selling costs and local and state government taxes to home buyers’ user-cost of capital. Australian Economic Papers, 35 (66), 60–73. Wood, G.A. (2001) Are there tax arbitrage opportunities in private rental housing markets? Journal of Housing Economics, 10 (1), 1–20.

References

317

Wood, G.A. & Kemp, P.A. (2001) The taxation of British landlords: an international comparison. Paper presented at Tenth Annual International Real Estate Conference (AREUEA). Cancun, Mexico, 6–8 May. Wood, G.A. & Maclennan, D. (1982) Search adjustment in local housing markets. In: Modelling Housing Market Search (ed. W.A.V. Clark), 54–80. Croom Helm, London. Wood, G.A. & Watson, R. (2001) Marginal suppliers, taxation and rental housing: Evidence from microdata. Journal of Housing Research, 12 (1), 91–114. Wood, G.A. & Tu, Y. (2001) Are there rent clientele groups among investors in rental housing? Paper accepted for presentation at 10th Annual International Real Estate Conference, May 6–8, 2001, Cancun, Mexico. Wood, G.A., Watson, R. & Flatau, P. (2001) Tax preferences, effective marginal tax rates and the incentive to invest in residential rental housing. School of Economics Working Paper, presented at 30th Annual Conference of Australia Economists, Perth, September 23–26. Wood, P.W. & Preston, J. (1997) Assessing Housing Need: a guidance manual. Scottish Ofﬁce Development Department, Edinburgh. Yates, J. & Flood, J. (1987) Housing Subsidies Study. Housing Research Council, Project Series No.160. Yates, J. & Whitehead, C.M.E. (1998) In defence of greater agnosticism: a response to Galster’s ‘Comparing demand-side and supply-side subsidies.’ Housing Studies, 13 (3), 415–23. Yates, J. & Wulff, M. (2000) Whither low cost private rental housing? Urban Policy and Research, 18 (1), 45–64. Zodrow, G. (1983) Local Provision of Public Services: The Tiebout Model After Twenty-Five Years. Academic Press, NY.

Index

Abraham, J.M. 76 academic environment see universities access-space model see trade-off model Adair, A. 54, 84, 195 Adam Smith Institute 272 Adams, D. 196 adjustment costs 59–60 administrative costs 58–9 administrative failure 13, 140, 148, 150–1 adverse selection 103 aesthetics 255 affordability 15, 42, 173, 182–8, 239 aggregation 104–6 Agpar, W. 31 Allen, F. 150 allocation of social and public housing 137, 148, 152 allowances 31 Alonso, W. 22, 23, 84 Ambrose, B.W. 77 Amemiya, T. 79 Anas, A. 23, 26, 30, 31, 76, 196 Andrew, M. 32, 105 Angel, S. 86 Anglin, P.M. 86 Anstie, R. 119 appraisal of housing 85 Apps, Patricia 251 arbitrage, taxation 12, 118–22 Arnott, R. 22, 23, 26, 30, 31, 196 Arrow, K.J. 140 asset pricing models 175 asymmetric information 140 Atkinson, A.B. 3 Attanasio, O.P. 93 Australia legal and administrative costs 59

research on housing in 251 search costs 58 social and public housing 137 taxation in 112, 114–15, 116, 117, 120, 122, 128 Austrian economics 2 Auten, G.E. 127 Awan, K. 74, 85 Baker, R. 150 Ball, M. 83, 95, 97, 193, 195 Ballas, D. 229 Baratz, G. 22, 29 Barclay, Peter 279 Barkham, R. 98 Barlow, J. 196 Barr, N. 138 Bartik, T.J. 60 Bartlett, W. 195, 196 Barton, D.M. 85 Bassett, K. 237 Basu, S. 86 Bate, R. 196 Beazley, M. 237 Bednarz, R. 85 Bender, A. 84 Berg, L. 94 Bergstrom, R. 94 Berry, B.J. L. 54, 85 Berry, J. 54 Best, R. 278 bidding 39, 40 Black, J. 98 Blackaby, B. 136, 221 Blackaby, D.H. 94 Blackaby, R. 224–5, 231 Boelhouwer, P.J. 93

320

Index

Bondﬁeld, Margaret 249 Bös, D. 145 Bourassa, S.C. 40, 51, 53, 82, 110, 112, 129, 130 Bover, O. 94 Bowden, R.J. 76 Bowley, Marian 249 Box, G.E. P. 81 Boyle, M.A. 84 Bramley, G. 193, 195, 196, 198, 200, 201, 204, 205, 206, 210, 214, 221, 222, 230 Bratt, R.G. 141 Bromwich, M. 150, 151 Brook, R. 221 Brown, J. 104 brownﬁeld land 203 Brueggeman, W.B. 116 Buchel, S. 85 Burgess, J.F. 83 Burns, L. 139 Burrows, R. 142 business cycles 64, 92 businesses housing and business start-ups 98 neighbourhoods and 155 Butler, R.V. 78 Butts, M.S. 85 Cameron, G. 94, 95 Cameron, S.J. 85 Can, A. 86 Canada search costs 57 taxation in 112, 116 capital asset market 39 Capozza, D.R. 80 careers, housing 255 Carr-Hill, R. 142 Case, K.E. 72, 94, 102 Cassel, E. 52 Central Housing Advisory Committee 256 Centre for Environmental Studies 256, 259 Centre for Policy Studies (CPS) 272 Chambers, D. 59 Champion, A. 201, 204 Chan, S. 60, 61 Chaplin, R. 151 Charity Organisation Society 256, 257 Charles, K. 173, 178, 181 Charles, S. 138 Cheshire, P. 84, 193, 195, 199 Cho, M. 102 choice of housing 2–3, 39, 42–3 tenure choice constraints 174–8 Chowhan, J. 84

Christensen, L.R. 80 Chun, G. 84 Chung, K.S. 54 cities see urban areas Clapham, D. 96, 240 Clapp, J.M. 85, 86 Clark, W.A. V. 57 Clarke, G. 229 Cliff, A.D. 51 Clotenfelter, C. 127 collective socialisation theories 164 Colwell, P.F. 74 complexity in housing markets 12, 106–7 construction of housing 96–7, 108, 193 research on 251, 255 consumers housing and expenditure by 92–4 neighbourhoods and 157–8 contagion models 164 conveyancing 59 Coulson, N.E. 254 Court, A.T. 74 Cox, D. 81 Crane, J. 164 Cressy, R. 98 Crosland, Anthony 136 crowding-out 97 Cullingworth, J.B. 196, 225, 249 Cunningham, D.F. 59 Dale-Johnston, D. 40 Daniell, J. 219 Davidson, R. 22 Davies, H.W. E. 225 Davis, O.A. 139 De Borger, B. 79 de Leeuw, F. 30, 31, 116 Deakin, Nicholas 259 DEMOS 272 Denmark 96 deregulation in ﬁnancial sector 4, 92, 93, 104–5 Des Rosiers, F. 84 Dilmore, G. 74 Din, A. 84 DiPasquale, D. 23, 40, 58, 59, 103, 228 disability 255 disequilibrium hedonic price models 75–6 housing market 41, 42–51 distributional issues social and public housing and 13, 143–7 taxation and 112–17 Donnison, David 249, 250 Doyal, L. 265

Index Dreiman, M. 76 Dubin, R.A. 85, 86 Duclaud-Williams, Roger 250 Dunham, J. 165 durability 38–9, 156, 175 dwelling ﬁltering 30 Dynarski, M. 60 Economic and Social Research Council 261, 276 economy economic cycles 64, 92 housing and 11–12, 91, 92–8, 107–8, 254 Edel, M. 85 edge cities 26 efﬁciency in housing market 102–3 neighbourhood change and social efﬁciency 165–7 provision of social and public housing and 13, 138–43 elderly people 255 Englund, P. 84, 102 Ensor, R.C. K. 249 environmental issues 84, 198–9, 255 Epple, D. 85 equity withdrawal 93 Erekson, J. 70, 71 Ermisch, J. 3, 86 ethical issues 6 Eum, S.J. 76 European monetary union 92, 95–6 Evans, A. 23, 193, 198, 199, 200, 205 Eversley, David 259 expectations 60–1, 93, 166 rational 102 Fabian Society 256, 272 Fair, R.C. 76 Fallis, G. 25, 31 Family Policy Studies Centre 280 Feldstein, M. 127 ﬁltering models 9, 22, 27–30, 36, 146 ﬁnancial sector, deregulation and liberalisation 4, 92, 93, 104–5 ﬁnancing costs 61 Fischel, W. 196, 199 Fisher, F. 74 Fisher, J.D. 116 Fisher, L.M. 254 Flatau, P. 131 Flood, J. 110 Follain, J.R. 67, 70, 75, 76, 77, 79, 82, 83, 84, 116, 119 Fordham, R. 221

321

forecasting house prices 12, 91, 98–107 Forrest, R. 236, 237, 238, 240, 241 Forrester, Jay 259 Foulis, M.B. 238 France 95, 96 social and public housing 137 Freeman, A.M. 84 Friedman, J. 59 Friedman, Milton 250 Fujita, M. 23, 25, 26 Gabriel, S.A. 82 Gale, D. 150 Galster, G.C. 29, 36, 38, 84, 107, 139, 144, 145, 157, 158, 162, 166, 167, 219 game theory 164–5 Garrod, G. 84 Gatzlaff, D.H. 78, 102 Genesove, D. 61 Germany 95, 96 social and public housing 137, 146 Giaccotto, C. 85 Gibb, K. 9, 23, 27, 30, 32, 36–7, 85, 229 Gilley, O.W. 85, 86 Glasgow citywide model 9, 32–5 Goldberger, A.S. 80 Goldﬁnch, J. 39, 42 Goodlad, Robina 230, 260 Goodman, A.C. 40, 41, 51, 53, 74, 85, 195 Goodman, J.L. 77, 78 Gordon, I. 95, 206 Gordon, R.H. 119, 126 Goss, S. 221 Gough, I. 265 governance, social and public housing 148–51 government and the state academics in Parliament 271–2 administrative failure 13, 140, 148, 150–1 advisers to 273–4 think tanks 256–7, 272–3 transaction costs and 65 see also policies; social and public housing Granovetter, M. 164 Grant, M. 196 Grant, R.A. 223 Gravelle, J.G. 126, 127 Grebler, L. 139 Greece 95, 96 Green, R.K. 64, 65, 72, 84, 129, 130 Greenﬁeld, Susan 272 Greer, S. 155 Greve, John 249, 250 Grey, A. 3 Grigsby, W. 2, 22, 27, 28, 29, 82, 110, 112, 130, 153, 157, 158, 159, 161

322

Index

Griliches, Z. 74 Gross, D.J. 83 Grossman, S. 175, 176 Guisanni, B. 195 Gyourko, J. 179, 181, 182, 183, 184, 186, 188 Haas, G.C. 74 Haavio, M. 64 Hadjimatheou, G. 195 Hahn, Frank 3 Hall, Peter 259 Hall, S. 104, 105 Halvorsen, R. 52, 79, 80, 81 Hamilton, B. 196 Hancock, K. 141 Hansmann, H. 150 Harmon, O.R. 83 Harrison, A.J. 139 Hasegawa, H. 84 Haurin, D.R. 78, 103 Hausman, J.A. 83 Hawtin, M. 221 Hayek, Friedrich 2 Heal, G.M. 230 Healey, P. 194 health, poor housing and 141–2, 254 hedonic price models 3, 11, 22, 52, 67–87, 199 conceptual issues 75–6 current position 82–3 dependent variable 77–8 equilibrium and disequilibrium 75–6 example of applications 83–5 functional form 79–82, 86 fundamental hedonic equation 68–9 hedonic price index 68–71 identiﬁcation problems 75 independent variables 78–9, 81–2 market and sub-market deﬁnition 82 repeat sales models 71–3 roots of 73–4 second stage models 69–71 single-equation models 82–3 speciﬁcation issues 76–82 Hendershott, P.H. 59, 64, 65, 76, 114, 116, 119, 126, 128, 129 Henderson, J.V. 23, 24 Henley, A. 93 Heseltine, Michael 236–7 Hess, E. 59 heterogeneity 39 Hicks, John 74 Hills, John 3, 144, 242, 254, 279 Hirsch, F. 2 Hirshleifer, J. 140

Hocking, R.R. 79 Hoesli, M. 84, 85 Hoffman, J. 84 Holmans, A. 30, 151 Holmes, C. 30, 130 Hölmstrom, B. 148 Holt, C.C. 127 homelessness 254, 262 Hopkins, C. 142 Hort, K. 93 households, neighbourhoods and 155 housing deﬁnition of 3 see also individual topics Hoyt, Homer 28 Hsiao, C. 200 Hu, S.C. 128 Hughes, G.A. 94–5, 236 Hunter, A. 155 Hurst, E. 173, 178, 181 income, ratio of house prices to 99–100 information 3, 7, 140 Ingram, G. 23, 30 Institute of Contemporary Scotland 265 Institute of Economic Affairs 257, 272 interest rates affordability and 182–3 expectations 60 sensitivity to 105 investment in housing 38–9 over-investment 111 rental housing 115–17 Ioannides, Y. 102 Ireland 95, 96 Italy 95, 96 Ittner, J.B. 77, 78 Jaffe, A.J. 59 Jaffee, D.M. 76 Janssen, C. 84 Jimenez, E. 67, 70, 75, 83 John Smith Institute 272 Jones, C. 54, 226, 228, 236 Jones, L.D. 172, 173, 175, 176, 181 Joseph Rowntree Foundation 269, 270, 274, 277, 279 Jud, G.D. 86 Kain, J. 31, 77, 84 Kaufman, D. 64 Kauppi, H. 64 Keeble, D. 98 Kellet, J. 142 Kemp, P.A. 116

Index Kennedy, P.E. 81 Kerr, M. 238 Kiefer, D.M. 118 Kiel, K.A. 84 Killen, S. 139 King, A.T. 83, 85 King, M. 3, 93, 128 Kirwan, R. 195 Kleinman, M.P. 145 Knight, J.R. 86 Koskela, E. 94 Kurz, C. 84 labour market 82 housing and 92, 94–5 Laidler, D. 111, 127, 128 Lambert, C. 193 Lancaster, Kelvin 73–4, 77, 154 land use regulation see planning (zoning) regulation Laroque, G. 175, 176 Leamer, E.E. 79 legal costs 58–9, 65 Leishman, C. 201, 210, 214 Lentz, G.H. 116 Leven, C. 193 Levine, D.N. 164 Lie, R. 196 life cycle models 102, 103, 108 Lind, R.C. 140 Ling, D.C. 116, 119, 126, 129 Linneman, P. 82, 173, 176, 177, 178, 179, 181, 182, 183, 184, 188 Litzenberger, R.H. 117, 118 Llewellyn-Davies 194 local authorities see planning of housing; planning (zoning) regulation; social and public housing Lowry, I. 28 lump sum payments 77 Lusht, K.M. 85 Lynn, P. 238 McBeath, J. 142 McClure, K. 116, 130 McCormick, B. 94–5, 236 McDonald, J. 23, 25, 30 McFadden, D. 46 McGill, G.A. 129 McGreal, W.S. 54 Macho-Stadler, I. 140, 148 MacKay, D. 9, 32, 85 Mackintosh, J.M. 249 Maclennan, Duncan 1–8, 9, 18, 20, 22, 23, 26, 27, 28, 29, 30, 36–7, 39, 40, 41, 42,

323

56, 57, 67, 75, 82, 83, 87, 90–1, 95, 105, 108, 110, 135, 138, 145, 151, 153, 158, 168, 189, 195, 196, 218, 221, 226, 227, 228, 231, 232, 240, 242, 243, 244, 249, 254, 259, 273, 274, 278, 279, 280, 281 MacNee, K. 238 MacNevin, A.S. 116 macroeconomy, housing and 11–12, 91, 92–8, 107–8, 254 Maddala, G.S. 78 Malatesta, P.H. 59 Malpass, P. 237, 240 Malpezzi, S. 72, 76, 77, 78, 79, 81, 82, 84, 85, 86, 130, 196, 200 Mankiw, N.G. 103 Manning, D.N. 94 Manning, Ian 251 market for housing 3 aggregation in 104–6 complexity 12, 106–7 efﬁciency 102–3 failure 138–43, 150–1, 228 planning (zoning) regulation and 193–217 structural change 104–6 volatility 97–8, 103 see also price of houses; sub-markets in housing; subsidies; taxation Marmot, M. 265 Marris, Peter 251 Mason, C. 86 Massey, Doreen 259 May, H. 196 May, Robert 272 Mayer, C. 201 Mayo, S.K. 83, 86 Meadows, Donella 259 media 257–8 Meen, Geoff 4, 9, 20, 32, 37, 93, 94, 97, 104, 105, 106, 193, 195, 214, 254, 280 Meese, R. 86 Megbolugbe, I.F. 182 Mendelsohn, R. 52 merit goods 141 Meyer, C.J. 61 Michaels, R. 195 MIGMOD model 201–16, 217 demand side 206–10 geographical units 205–6 key variables 202–4 results 206–16 setting up 201–6 supply side 210–16 theoretical interpretation 204–5 migration, housing and 92, 94–5 Miles, D. 93

324

Index

Mills, E. 23, 82, 84, 128 Minford, P. 236 Mises, Ludwig von 2 mismatch 40, 41 mobility 10–11, 63–4 modiﬁcation of housing 39 monetary policy 97–8 monetary union 92, 95–6 Monk, S. 94, 194, 196, 199, 226 monocentric residential model 23 More, A. 226 mortgages costs of 61 European comparisons 95–6 ﬁnancial sector deregulation and liberalisation and 92, 93, 104–5 Moser, Claus 272 Moulton, B.R. 84 moving costs 59–60 Mozolin, M. 84 Muellbauer, John 4, 91, 94, 95, 195, 254 Mulgan, Geoff 272 Munro, M. 195 Murie, A. 142, 236, 237, 238, 240, 241 Murphy, A. 94, 195, 254 Murray, Charles 250, 265 Musgrave, R.A. 141 Muth, Richard 23, 26, 84, 195, 250 Needham, B. 196 Needleman, Lionel 249 negative equity 93 neighbourhoods 13–14, 153–69 change 2, 14, 160–7 external forces and 163 macro process 160–3 micro processes 160, 163–5 social efﬁciency and 165–7 creation of 158–9 deﬁning 153–5 idiosyncrasies of 156–8 policy implications 167 regeneration 167 neoclassical economics 90, 100, 103, 105, 108 Netherlands consumer expenditure and housing in 92, 93 debt in 96 social and public housing 137, 147 subsidies in 146 Neutze, Max 198, 251 Nevin, B. 142 Nevitt, Adela 249 new urban economics 23 New Zealand, taxation in 112

Nicholson, M. 85 Niskanen, W.A. 230 noise in markets 3 Nordvik, V. 119 Nourse, H.O. 77 Nozick, Robert 250 Oates, W. 85 Odling-Smee, J. 27, 42 Ohta, M. 82 Olsen, E.O. 22, 28, 29, 83, 85 O’Sullivan, A.J. 3 Oswald, Andrew J. 11, 64, 129, 254 owner-occupied housing access to home ownership in USA 14–15, 172–89 affordability 15, 173, 182–8 constraints to tenure choice 174–8 racial differences 14–15, 173, 178–81, 189 wealth constraints 14–15, 173, 176–7, 179–81 European comparisons 95 mobility and 10–11, 63–4 taxation and 111, 112–15, 118, 127, 128, 129 transaction costs 58, 59, 61, 63–4 Ozanne, L. 78, 84, 116 Pace, R.K. 85, 86 Pahl, Ray 259 Palm, R. 51 Palmquist, R. 80 Paris, C. 136 Pasha, H.A. 85 paternalism 148 Pawson, H. 238, 240 Pérez-Castrillo, D. 140, 148 Phillips, R.S. 77 Pines, D. 22 planning of housing 16, 218–32 land use planning and 225–6 policy and institutional framework 219–22 in practice 222–5 roots of problem with 226–32 planning (zoning) regulation 193–217 economic impact 198–201 housing planning and 225–6 housing supply and 15, 194, 195–6, 210–16 inter-urban panel model 201–16, 217 demand side 206–10 geographical units 205–6 key variables 202–4 results 206–16 setting up 201–6 supply side 210–16

Index theoretical interpretation 204–5 regimes 196–8 Plaut, S.E. 176 Podogzinski, H.U. 199 policies ﬁltering models and 28 monetary policy 97–8 neighbourhood dynamics and 167 planning of housing 219–22 practice and 4–8 Right to Buy (RTB) 4–5, 8, 16–17, 235–46 evidence base for 1980 legislation 237–8 modernised RTB in Scotland 242–5 policy objectives 236–7 relationship between policy and research 241–2 research on 238–40 social and public housing 135–6 transaction costs and 10–11, 63–5 universities and 17–18, 268–81 crossing the divide 271–5 making connections 275–7 persistence and partnership 280–1 presentation and dissemination 277–80 separate worlds 268–71 Policy Studies Institute 280 Pollack, R. 150 Pollakowski, H. 52, 79, 81 Posner, R.A. 148 Potepan, M.J. 61 Power, Anne 260 predicting house prices 12, 91, 98–107 preference models 165 presentation 277–80 pressure groups 7–8 Preston, J. 221 price of houses 3, 38–9 affordability 15, 42, 173, 182–8, 239 change in 2 consumer expenditure and 92–4 expectations 60 ﬁltering and 29–30 forecasting 12, 91, 98–107 indices 22, 68–71, 84 macroeconomy and 11–12, 91, 92–8, 107–8, 254 neighbourhoods and 156–7 premiums 40 random walk hypothesis 102–3 ratio to income 99–100 see also hedonic price models prisoners’ dilemma 164 privatisation 150–1 product groups 2 Prud’homme, M. 84

325

Pryor, J. 279 psychic costs of moving 60 public housing see social and public housing quality of housing 29–30, 41, 42, 184 substandard housing 139, 141–2 Quigley, J.M. 23, 25, 28, 39, 60, 61, 77, 84, 85, 86, 140 racial differences 14–15, 173, 178–81, 189 Ramazan, G. 86 random walk hypothesis 102–3 rational expectations 102 Redfearn, C.L. 42 redistribution, social and public housing and 13, 143–7 regeneration programmes 142–3, 167 rent controls on 144 social and public housing 137, 152 value and 77 rent seeking 148, 193 rented housing 175 research on 255 taxation and 111, 115–17, 118–22, 129, 130–1 clientele effects 12, 124–5 lock-in effects 12, 125–7 transaction costs 57–8, 59, 65 see also social and public housing repeat sales models 71–3 research on housing 17, 248–66 changing environment for 255–62 academic changes 259–61 political changes 256–8 recent developments 261–2 signiﬁcance of changes 262 housing supply 195–6 policy and 268–81 presentation 277–80 priorities for Scotland 265 recent 251–5 research industry 262–5 resumption after WWII 249–51 Right to Buy (RTB) and 16–17, 238–40, 241–2 think tanks 256–7, 272–3 Richmond, P. 238 Right to Buy (RTB) 4–5, 8, 16–17, 235–46 evidence base for 1980 legislation 237–8 modernised RTB in Scotland 242–5 policy objectives 236–7 relationship between policy and research 241–2 research on 16–17, 238–40

326

Index

risk aversion 140 Robertson, D.S. 223 Robinson, Joan 1 Robinson, R. 3, 28 Robson, Brian 259 Rodgers, B. 279 Rohe, W. 158, 159 Rosen, H.S. 60, 110, 128, 174 Rosen, K. 25, 31, 128 Rosen, Sherwin 69, 73, 74, 77, 79 Rothenberg, J. 29, 30, 36, 41, 51, 53, 54, 82, 139, 161, 162, 229 Rothschild Report 261 Rouwendal, J. 85 Rowntree, Joseph 270 Rowntree, Seebohm 249 Russia 86, 250 Rydin, Y. 196 Sandercock, Leone 251 Sass, T.R. 199 Satsangi, M. 85 Savage, W.G. 249 savings 177 Schelling, T.C. 164 Schnare, A. 40, 53, 195 Sclar, E. 85 Scott, Suzie 260 Scottish Council Federation 265, 272 Seaks, T.G. 86 search activity 39–40, 41, 42–3 search costs 57–8, 65 segmentation in housing market 9–10, 38–42, 161 disequilibrium and 41, 42–51 identiﬁcation 10, 51–4 self-fulﬁlling prophecies 166 Sewel, J.B. 238 Shell, K. 74 Shelton, J.P. 127 Sheppard, S. 67, 75, 84, 195, 199 Shiller, R.J. 72, 102 Shilling, J.D. 85, 128 Short, J. 196 Short, K. 84 Simenauer, R. 82 Simonson, J. 59 Simpson, D.G. 230, 231 simulation models 9, 22–3, 30–2, 36 Glasgow citywide model 9, 32–5 Slemrod, J. 114 Small, K.A. 23, 26 Smith, G.R. 143 Smith, L. 25, 31 Smith, M.E. H. 220, 221

Smith, V.K. 195 social problems 164 social and public housing 3, 7, 13, 32, 135–52, 221, 254 economic analysis in 135–8 efﬁciency reasons for provision 13, 138–43 governance 148–51 privatisation 150–1 redistribution and 13, 143–7 Right to Buy (RTB) 4–5, 8, 16–17, 235–46 evidence base for 1980 legislation 237–8 modernised RTB in Scotland 242–5 policy objectives 236–7 relationship between policy and research 241–2 research on 16–17, 238–40 Soderberg, B. 84 Solow, R. 23 Somerville, C.T. 30, 130 Somerville, T. 201 Soong, R. 164 Sosin, H.B. 117, 118 Spain 95, 96 spatial aspects 2 immobility 38 neighbourhoods 155 topographical sub-markets in housing 40–1, 42–51 spatial models 22 stamp duty 59 Stein, J. 61 Stephens, Mark 4, 95, 96, 137 Stone, E. 142 Straszheim, M. 23, 26, 27, 41, 82 Stretton, Hugh 251 structural change in housing market 104–6 Struyck, R. 30, 31, 40, 53, 82, 195 sub-markets in housing 2, 22, 54–5, 57, 90–1 hedonic pricing model and 82 identiﬁcation 51–4 model of disequilibrium and 42–51 neighbourhood change and 160–1 quality and 29–30, 41, 42 segmentation 9–10, 38–42, 51–4, 161 disequilibrium and 41, 42–51 topographical 40–1, 42–51 subsidies 3, 6, 13, 28, 31–2, 144–6 see also taxation substandard housing 139, 141–2 Suits, D.B. 81 Sumka, H. 70, 71 supply of housing 15, 194, 195–6, 210–16 Sutherland, Stewart 272 Sweden social and public housing 137

Index subsidies in 146, 147 Sweeney, J. 30 switching regression model 76 Switzerland 137 Taub, R.D. 165 taxation 12, 110–31, 155 arbitrage 12, 118–22 capitalisation into house prices 85 concessions 3 distributional issues 112–17 expectations 60 owner-occupied housing and 111, 112–15, 118, 127, 128, 129 rented housing and 111, 115–17, 118–22, 129, 130–1 clientele effects 12, 124–5 lock-in effects 12, 125–7 transaction taxes 59 welfare losses 12, 127–9 Taylor, G. 165 Taylor, Laurie 260 technical progress 96 Temkin, K. 158, 159 tenure choice constraints 174–8 Tewdyr-Jones, M. 196 Thatcher, Margaret 150, 235, 272 theory, perspectives on 1–4 Theriault, M. 84 Thibodeau, T.G. 40, 41, 51, 53, 85, 86 think tanks 256–7, 272–3 Thomas, A. 95 Tiebout, C. 85, 196 Tirtiroglu, D. 102 Titman, S.D. 118 Titmuss, Richard 250 Tiwari, P. 84 topographical sub-markets in housing 40–1, 42–51 town planning see planning (zoning) regulation Toynbee Hall 256, 257 Tracy, J. 182, 183, 186, 188 trade-off models 9, 23–7, 35–6 basic trade off 23–6 criticisms 27 evidence 26 extensions 23, 26 training 8, 260 transaction costs 10–11, 40, 56–65, 149 adjustment costs 59–60 expectations and uncertainties 60–1 ﬁnancing costs 61 legal and administrative costs 58–9, 65 model 61–3

327

policy implications 10–11, 63–5 search costs 57–8, 65 taxonomy 57–61 transport 251 Triplett, J.E. 74 Troy, Patrick 251 Truesdale, D. 238 Tsoukis, C. 195 Tu, Y. 39, 41, 42, 48, 50, 82, 90, 108, 116, 117, 124, 125, 130, 195, 226 Turner, B. 85, 137, 147 Twine, F.E. 238, 240 uncertainties 60–1 unemployment 10–11, 63–4, 65, 94 United Kingdom affordability in 189, 239 consumer expenditure and housing in 92, 93–4 ﬁnancial deregulation in 4, 104–5 housing economics in 1 legal and administrative costs 59 migration within 94–5 planning of housing in 16, 218–32 planning (zoning) regulation in 193, 194, 196, 197–8, 199 inter-urban panel model 201–16, 217 policy issues in 4–5, 6–7 research on housing in 249–50, 251, 254–66 simulation models in 31, 32–5 social and public housing 3, 32, 135–6, 143, 147, 151–2, 221 Right to Buy (RTB) 4–5, 8, 16–17, 235–46 structural change in housing market 104–6 subsidies in 3, 6, 146 substandard housing 139 taxation in 112, 128 United Nations 86 United States of America access to home ownership in 14–15, 172–89 affordability 15, 173, 182–8 constraints to tenure choice 174–8 racial differences 14–15, 173, 178–81, 189 wealth constraints 14–15, 173, 176–7, 179–81 adjustment costs 59–60 housing planning 218, 219–20 legal and administrative costs 59 neighbourhoods in 161 planning (zoning) regulation in 196, 199, 216 repeat sales model 72

328

Index

research on housing in 195, 250, 259 search costs 57–8 simulation models in 30, 31 social and public housing 137, 147, 240 subsidies in 145–6 taxation in 112, 115–16, 127–8, 130 unemployment 64 universities 249, 259–61, 263 policy and 17–18, 268–81 crossing the divide 271–5 making connections 275–7 persistence and partnership 280–1 presentation and dissemination 277–80 separate worlds 268–71 urban areas 2 hedonic pricing model in 84 housing models in 22–37 ﬁltering models 9, 22, 27–30, 36 Glasgow citywide model 9, 32–5 simulation models 9, 22–3, 30–2, 36 trade-off models 9, 23–7, 35–6 inter-urban panel model 201–16, 217 sub-markets in housing 51–4 see also neighbourhoods Urban Institute (UI) model 31 value, rent and 77 Van Zilj, V. 221 Vandell, K.D. 84, 111, 129, 130, 161 Varady, D.P. 222–4 Venti, F. 60 Viren, M. 94 volatility in housing market 97–8, 103 Wachter, S. 173, 176, 177, 178, 179, 181 Waddell, P. 54 Waddilove, Lewis 249 wages, housing and 92, 94 Walker, S. 98 Wallace, H.A. 74 Wallace, N. 86 Watkins, C. 51, 53, 54, 200, 201, 204, 210, 226, 238, 240

Watson, R. 117, 130, 131 Watts, Harold 81 wealth 172, 175 constraints 14–15, 173, 176–7, 179–81 Webb, Steve 271 Weber, G. 93 Weber, M. 164 Weibull, W.J. 45 Weicher, J.C. 129 Weil, D.N. 103 Weinberg, D.H. 59 Weinberg, D.J. 58, 59 welfare economics 1 welfare losses in taxation expenditures 12, 127–9 Westaway, P. 195 Wheaton, W. 23, 40, 58, 59, 103, 228 Whinston, A.B. 139 Whitehead, Christine 27, 42, 83, 136, 137, 138, 141, 144, 145, 146, 147, 150, 194, 196, 226, 274 Wilkinson, R. 142, 265 Willetts, David 272 Williams, N.J. 235, 236, 238, 240 Williamson, Oliver 2, 3, 148, 149 Willis, K.G. 84, 85 Willmott, Peter 251 Wilson, W. 236 Wiltshaw, D.G. 198 Wise, D.A. 60 Witte, A.D. 70, 71, 74, 85 Wood, A. 95, 97 Wood, Gavin 2, 3, 40, 42, 57, 58, 112, 116, 117, 124, 125, 128, 130 Wood, P.W. 221 World Bank 83 Wulff, M. 117, 130 Yates, J. 110, 117, 130, 144, 145, 146 Young, Michael 251 Zodrow, G. 85 zoning see planning (zoning) regulation